Cotton in Africa

F a c t sS e r i e s

Table of contentsAbstract 4Facts and figures 61. Cotton, an important cash crop for African smallholder farmers 72. Threats to the sustainability of cotton production in Africa 123. Bt cotton to limit crop losses due to insect attacks 16Development of Bt commercial cotton cultivars and hybrids 20Production of Bt cotton cultivars and hybrids and their seed provision 21Commercialization of Bt cotton in Africa is spreading more slowly than in other cotton growing regions 22What happened with Bt cotton in Burkina Faso? 254. Lessons learned from experiences in early-adopting Bt cotton countries worldwide 28Changes in yield and insecticide use 29Technology costs need to be considered and their impacts can be variable 32Cotton is a demanding crop in terms of crop husbandry and pest management 34 Btcottonbenefitsdependonpestabundance 34 Consequencesofreducedinsecticideusefornon-targetpests 34 Performanceundermarginalgrowingconditions 35 Sustainabilityofthetechnology-howtomanagetheevolutionofinsectresistance toBtcotton 365. The impact of agricultural policies on cotton production 40How the introduction of Bt technology has affected the organization of the input delivery 41Institutional structures governing cotton production in Sub-Saharan Africa 43Sector structure has a major influence on performance 45Farmers’ access to information on managing the Bt technology 466. Conclusion 487. References 49

Source: Bruno Tinland, personal collection 3

Cottonrepresentsacrucialsourceof income in

Sub-Saharan Africa, both for rural populations

andfornationaleconomies.Itisoneofthemost

widely produced cash crops grown by African

smallholderfarmers,rankingonlysecondinvalue

after cocoa. Despite its economic potential, the

cotton industry is subject to a number of risks,

such as price fluctuations of both inputs and

cotton on the world market, changing weather

conditions, pest attacks and also problems

relatedtopestsbecomingresistanttopesticides.

Alltheseriskfactorsthreatenthesustainabilityof

cottonproductioninAfrica.

InmostSub-SaharanAfricancountries,yieldsof

500-700 kg/haof seed cottonproducedunder

rainfed conditions are typical for varieties with

ayieldpotentialcloseto3,000kg/ha.Cotton is

subject to damage by an extraordinarily wide

range of insect pests, which causes significant

losses to cotton production and impacts fiber

quality.The larvaof thecottonbollworm is the

main cotton pest throughout Africa, which can

cause damage in up to 90%of bollswhenun-

treated.Duetostrongbollwormpestpressure,

cotton is heavily sprayed with chemical pesti-

cides. This poses significant health hazards for

many farmers and laborers and generates ex-

tensiveenvironmentalpollution.

In 2014, cottonwas planted on 35million hec-

tares globally, 64% (22.3 million hectares) of

which was insect-resistant genetically modified

cotton, referred to as Bt cotton. In 2016, a to-

talof8Africancountrieseitherplanted,actively

evaluated field trials or moved towards grant

approvals for the general release of Bt cot-

ton. South Africa was the first country on the

AfricancontinenttoadoptBtcottonforcommer-

cialproductionin1998,followedbyBurkinaFaso

in2008andSudan in2012. In thesecountries,

Bt varieties spread rapidly because they provi-

dedasignificantreductionofinsecticideuseand

bollwormdamage,anincreasedyieldandhigher

farmerprofits,andtheyallowedconservationof

beneficialnaturalenemies.

Abstract

However, in 2016, Burkina Faso government

temporarily suspended the growing of Bt cot-

ton to address a concern about fiber length

observedwith the varieties farmershavegrown

overthelasteightyears.Thisdecisionshowsthe

important role that technology developers and

breeders must play in incorporating traits and

qualitieswell-adaptedtolocalconditionsinorder

tomeetfarmerandmarketneeds.

In Africa, smallholder cotton growers often lack

access toproductivity-enhancing inputs suchas

improved seed, fertilizers, water and informa-

tion.Thecreditneededtofinanceinvestmentin

these inputs is themajor constraint. The imple-

mentation of reforms that are in linewith local

realities that cotton farmers face is required to

enhance efficiency in cotton production and to

revitalize Sub-Saharan African cotton sectors,

while inducing economic growth and alleviating

poverty. Donors and governments that invest

in the hope that genetically modified crops will

bringsignificantimprovementstothelivelihoods

of resource-poor farmers without first paying

attention to the fundamental institutions that

support broader agricultural development and

technologygeneration(withorwithoutgenetical-

lymodifiedcrops)are indangerofmisallocating

theirresources.Thelocalkeyinstitutionsinclude

those that support public and private capacity

for technology generation, technology delivery

throughmarkets,extensionandregulations,and

farmercapacitiestodemandservices,participate

inmarketsandunderstand the technology they

areusing.

5

Facts and figures Cotton is predominantly a smallholder crop, mainly grown on small family farms of less than 4 hectares in size, and is a crucial source of cash income (60%) for millions of farmers and their families in more than 20 countries across all regions of Sub-Saharan Africa.

Among export crops with substantial smallholder farmer involvement in Sub-Saharan Africa, cotton ranks second in value after cocoa.

During the decade of 2004-2014, the African continent contributed 6% to the world’s total seed cotton production (world production was about 1.4 million metric tons).

In most Sub-Saharan African countries, yields of 500-700 kg/ha of seed cotton produced under rainfed conditions are typical for varieties with yield potential close to 3,000 kg/ha.

The larva of the cotton bollworm is the main cotton pest throughout Africa, causing damage in up to 90% of bolls when untreated, leading to lost cotton production.

Half of the insecticides used in Africa are sprayed onto cotton, posing significant health hazards for many farmers and laborers and causing extensive environmental pollution.

In 2014, cotton was planted on 35 million hectares globally, 64% (22.3 million hectares) of which was insect-resistant genetically modified cotton, referred to as ‘Bt’ cotton.

In 2016, a total of 8 African countries either planted, actively evaluated field trials or moved towards grant approvals for the general release of Bt cotton (Cameroon, Ethiopia, Kenya, Malawi, Nigeria, South Africa, Sudan, Swaziland).

South Africa was the first country on the African continent to adopt Bt cotton for commercial production in 1998. Burkina Faso was the second country to adopt Bt cotton in 2008, and Sudan became the third in 2012.

In 2016, the government of Burkina Faso temporarily suspended the growing of Bt cotton to address a concern about fiber length observed in the varieties farmers have grown over the last eight years.

Smallholder cotton growers often lack access to productivity-enhancing inputs such as improved seed, fertilizers, water and information, and credit needed to finance investment in these inputs is the major constraint.

The implementation of reforms that are in line with local realities that cotton farmers face is required to enhance efficiency in cotton production and to revitalize Sub-Saharan African cotton sectors while inducing economic growth and alleviating poverty.

Cotton in Africa

Cotton, an important cash crop for African smallholder farmersCotton and cotton textile industries are central to the economic growth of both developed and developing countries. Cotton cultivation has been an important economic activity for thousands of years across all continents. Cotton occupies a relatively small percentage (3%) of the world’s crop acreage, a fact that may seem at odds with the prevalence of cotton in home textiles and clothes. However, cotton is one of the most significant crops in terms of land use after food grains and soybeans (FAOSTAT, 2017).

1

7Source: Bruno Tinland, personal collection

Marketing

Cotton is a perennial plant that has been do-

mesticated to grow as an annual crop. Natural

acclimatization processes have impacted cotton

throughout itshistory,butexactlywhen thedo-

mestication process began is unknown. Cotton

is grown around the world, from the tropics to

latitudeshigherthan40°(UzbekistanandXinjiang

Province inChina) (Figure1.1),eitherasdryland

(reliantonrainfall)orasirrigatedcotton(requiring

supplementalwatersupply).Thebasicconditions

requiredforthesuccessfulproductionofcotton

include a long frost-free period, a temperature

range of 18–32° C and 600–1200mmof water

overthegrowingcycle,whichtypicallylasts125–

175 days (FAO, 2012). Cotton exhibits a certain

degreeoftolerancetosaltanddroughtandcan

thereforebegrowninaridandsemi-aridregions.

However, higher and consistent yield and fiber

quality levels are generally obtained with irriga-

tionorsufficientrainfall.

Figure 1.2: Schematic representation of the cotton textiles value chain (source: adapted from FAO and ICAC, 2015 and reproduced with permission).

Figure 1.1: Cotton-producing countries (seed cotton, 2004-2014 average) (source: adapted from FAOSTAT, 2017 and reproduced with permission).

<= 1200

Tons

<= 17000

<= 85000

<= 322000

> 322000

EUROPE

AFRICA

ASIA

AUSTRALIA

SOUTH AMERICA

NORTH AMERICA

Final Processing

Designing

Finishing

Weaving/Knitting

Dyeing

Spinning

ProcessingProduction

Farm level

Ginning

Growing

Purchase of inputs

Cotton in Africa

A

B

C

Figure 1.3: (A) Burkinabe women picking cotton bolls (source: Bruno Tinland, personal collection); (B) spinning cotton into thread by hand; (C) transforming yarns into fabric by weaving.

Cottonhashundredsof applications, fromblue

jeanstosoap.Mostpartsoftheplantareuseful,

themostimportantbeingthefiberorlint,which

isusedtomakeyarnforcottoncloth.Thelinters

(theshortfuzzontheseed)provide,amongother

products,celluloseformakingplastics,explosives

andothergoods.Thecottonseeditselfiscrushed

intothreeseparateproducts:oil,mealandhulls.

Theoilisprimarilyusedforcookingoilandsalad

dressing; the meal and hulls that remain are

mainlyusedeitherseparatelyor incombination

aslivestock,poultryandfishfeed,andalsoasfer-

tilizer.Theprotein-richseedscanbeusedasfeed

for ruminants. Cottonseed contains gossypol,

a polyphenolic aldehyde, which can make cot-

tonseed toxic tomonogastric animals (e.g. pigs,

poultry,fish).Inruminants(e.g.cattleandsheep),

itisunlikelythatenoughcottonseedmealwould

beingestedtoresultintheanimalsufferingfrom

gossypoltoxicity(Gadelhaet al.,2014).

As textile industries grew and became more

specialized during the Industrial Revolution at

Cotton issimultaneouslyanagriculturalproduct

and industrial raw material. The cotton market

isoneofthemostdiversecommodityindustries

worldwide and the processing, transport and

by-product industries create significant oppor-

tunitiesforemploymenttohundredsofmillions

offarmersandprocessors(Kooistraet al.,2006).

The value chain begins with the farmer, who

growsandharvestsseedcottonfromthebollsof

theplant.Cottonproductionsystemsrangefrom

labor-intensivesystemsinAfricaandAsiatohighly

mechanizedsystemsintheUnitedStates,Australia

andBrazil. Byweight, seed cotton is composed

of roughly one-third cotton lint and two-thirds

cottonseed. The cotton lint is separated from

the cottonseed (a process known as ginning)

usingacottongin.Cottonlintisthensoldtospin-

ners who produce yarn. Textile manufacturers

transform yarns into fabric by knitting orweav-

ing theyarnsandapplyingdyesandfinishes. In

the final stage, end products (garments, home

textiles etc.) aremade from fabrics (Figures 1.2

and1.3).

9

the end of the 18th century, increasing demand

was placed on cotton production worldwide. In

Africa,thedemandforcottonwasalsoaprimary

drivingforceofmanycolonialregimes(Isaacman

and Roberts, 1995). This often led to the en-

forced cultivation of cotton and the disruption

of traditional economic and farming systems.

Today, although continually threatened by syn-

theticfibers,demandforcottonremainsstrong.

Withatotalproductionof21millionmetrictons

of cotton fiber lint during the 2015-2016 grow-

ing season acrossmore than 75 countries (the

majority indevelopingcountries), thesocial and

economicimportanceofcottononaglobalscaleis

self-evident(ICAC,2017).

Cottonispredominantlyasmallholdercrop.It is

grownmainly on small family farms less than 4

hectaresinsize(Gouseet al.,2003).From2004-

2014, the African continent contributed 6% to

world seed cotton production (world produc-

tion was about 1.4 million metric tons) (Figure

1.4) (FAOSTAT, 2017). Cotton is a major source

of foreign exchange earnings in more than

20 countries across all regions of Sub-

Figure 1.4: Production share of seed cotton by region (average 2004-2014) (source: adapted from FAOSTAT, 2017 and reproduced with permission).

A

Saharan Africa and a crucial source of cash

income formillions of smallholder farmers and

theirfamilies.Therefore,thecropiscriticalinthe

fightagainstruralpoverty.

BurkinaFasoandMaliarebyfarthebiggestcotton

producers in Africa (Table 1.1). In Burkina Faso,

more than 2millionBurkinabe citizens derive a

majority of their income (60%) from producing,

ASIA 69,4%

AMERICAS21.1%

EUROPE 1.5%

OCEANIA 2.1%

AFRICA 6%

Cotton in Africa

B

Figure 1.5: (A) Field cotton and (B) harvested cotton in Burkina Faso (source: Bruno Tinland, personal collection).

Table 1.1: Leading cotton producers in Africa in 2014

(source: FAOSTAT, 2017).

Country Production (tons of lint)

Yield (kg/ha of seed cotton)

Benin 102,600 941

BurkinaFaso 256,500 1374

Cameroon 81,000 1250

Chad 48,000 415

Egypt 113,000 3386

Ethiopia 38,000 951

Ghana 5,600 875

Guinea 15,000 977

IvoryCoast 132,000 976

Kenya 4,450 554

Malawi 53,800 644

Mali 232,748 1017

Mozambique 33,000 618

Nigeria 105,000 682

Senegal 11,000 1060

SouthAfrica 8,741 2954

Sudan 95,000 2683

Swaziland 600 508

Tanzania 81,000 790

Togo 37,000 815

Uganda 26,600 1056

Zambia 39,700 963

Zimbabwe 75,000 507

ginning,ortransportingcotton(Figure1.5).Public

servicessuchasschools,roads,publichealth,and

a variety of agricultural extension services have

traditionallybeenprovidedviacottonrevenues.

Moreover, studies in Ivory Coast and Zimbabwe

haveshownapositivesynergybetweencottoncul-

tivation,ontheonehand,andfoodproductionand

householdnutrition,ontheother.Thisisbecause

the income and experience gained through cot-

toncultivationareappliedtotherestofthefarm

(Sahn, 1990; Govereh and Jayne, 1999). In West

Africa, many cotton-producing regions have also

experienced significant growth in cereal produc-

tionbecauseofthedevelopmentofinputprovision

(i.e. energy, water, fertilizers and pesticides) and

the emphasis on agricultural innovation (OECD,

2006).Amongexportcropswithsubstantialsmall-

holderfarmerinvolvementinSub-SaharanAfrica,

cottonrankssecond invalueaftercocoa,and its

productionisspreadacrosstheAfricancontinent

(FAOSTAT,2017). Insomecountries,especially in

theSahel,thereisnoshort-tomedium-termcash

cropsubstitutetocottonforsmallholderfarmers.

11

Threats to the sustainability of cotton production in AfricaDespite its economic potential, the cotton industry is also subject to a number of risks, such as input price fluctuation, cotton price fluctuations on the world market, changing weather conditions, pest attacks and problems related to pests becoming resistant to pesticides. All these risk factors threaten the sustainability of cotton production in Africa (Vitale and Greenplate, 2014).

2

Source: Bruno Tinland, personal collection

Byfarthemostvisibleissuerelatedtocottoncul-

tivation,particularlyinAfricancountries,hasbeen

therelationshipbetweenlowinternationalprices

and domestic support of cotton production. To

be specific, subsidiesofferedby theUSA,China

andtheEuropeanUniontotheircottongrowers

cause significant market distortions. In 2003, a

groupofWestAfricancotton-producingcountries

(Benin,BurkinaFaso,ChadandMali)hasfileda

case with the World Trade Organisation asking

fortheremovalofproducersupportbytheUSA,

ChinaandtheEuropeanUnion(Baffes,2011).The

issuehascapturedpublicattentionasanexample

ofhowprotectionofagriculturebyindustrialized

countries can affect the livelihoods of produc-

ers in developing countries. A study in Benin

estimatedthata40%drop incottonprices (like

theonethatoccurredin2002)causedan8%rise

inruralpoverty,wherecottonaccountedfor22%

ofthegrossvalueofcropproductioninBenin(Mi-

notandDaniels,2005).

Cottonissubjecttodamagebyanextraordinarily

widerangeofinsectpests,whichcausessignifi-

cant losses to cotton production and impacts

fiberquality.Asextensiveareasof cottonhave

beenplanted,thecallforhigheryieldsandmore

uniformfiberresultedinmonocultures,increas-

ingcotton’ssusceptibilitytopestsanddiseases.

Pests are a significantproblem in Sub-Saharan

Africa,withclimaticconditionsfavoringmultiple

pestgenerationsperyearandheavypestdensi-

ties. About 15%of cottonproducedworldwide

is lost due to insect attack every year (Oerke,

2006). InWest Africa, the numbers are higher,

with about 25-35% of cotton lost because of

insects (Vitale et al., 2016). Among insects, the

larvaofthecottonbollwormisthemaincotton

pest throughout Africa. It has been reported

to damage up to 90% of bolls when untreat-

ed (Vitale and Greenplate, 2014) (Figure 2.1).

Otherbollwormsvaryfromcountrytocountryand

includepinkbollworm,tobaccobollworm,spotted

and spiny bollworms, and red bollworm. Dam-

agetocottonplantsischaracterizedbyfeeding

activity on squares (flower buds), flowers and

cottonbolls.Theseformsofdamagearethemost

severe,astheydestroytheplant’sreproductive

parts anddrastically reduce yield (Abhiyanand

Abhiyan,2012).

BA

Figure 2.1: (A) Larva of the cotton bollworm (source: Bruno Tinland, personal collection); (B) larva of the pink bollworm

(source: Peng Wan, personal collection).

Large-scale cotton production also has serious

environmental consequences. The expansion of

cottonin19thcenturyUSAledtowidespreadero-

sion and soil exhaustion (Stoll, 2002). Although

the widespread use of chemical insecticides

for cotton throughout the world has helped

farmers producing higher yields, it is also the

source of many problems, particularly in devel-

oping countries and more specifically in Africa.

About a quarter of the world’s insecticide use

isdevotedtocottonandhalfof the insecticides

in developing countries are used on cotton.

Severalofthoseinsecticidesareclassifiedbythe

WorldHealthOrganisation as ‘highly hazardous’

(Kooistraet al.,2006).InAfrica,cottonproducers

sprayaboutsixtimesperyear,althoughasmany

CASE STUDIES IN CHINA - THE BEWILDERING ARRAY OF COMMERCIAL INSECTICIDE PRODUCTS FROM WHICH FARMERS MUST CHOOSEIn China, the insecticide market consists of many small, local companies. Until mid-2008, China only recog-

nized process-based and not substance-based patents (molecules or any active compound). Thus, manufacturing

products based on molecules patented elsewhere has been common. The large number of formulators led to a

plethora of products (often mixtures of two or more active ingredients) and farmers face a pesticide market that

includes thousands of trade names (Tripp, 2009). Despite the fact that state regulatory agencies monitor the mar-

ket, there are concerns about the quality of many available products (Meng etal., 2006). A survey of 150 farmers

in Shandong Province recorded 448 different pesticide products (of which the vast majority are insecticides) used

in Bt cotton fields, many of which were not officially registered. In 15% of the cases, researchers were even unable

to identify the active component (Pemsl, 2006). Thus, it is a challenge for farmers to distinguish between trusted

and fraudulent input sources.

A

Figure 2.2: (A) Field cotton of a smallholder farmer in Burkina

Faso (source: Karim Maredia, personal collection); (B) harvested

cotton being piled up in traditional reed stockage in Benin.

Cotton in Africa

astensprayingscanberequired.Moreover,the

availability and promotion of cheap, low-quality

insecticides combined with sub-optimal agri-

cultural practiceshave led to the emergenceof

insecticide resistance in anumberof pests and

thedeclineordisappearanceofnaturalenemies

ofcottonpeststhatformerlyhelpedinmaintain-

inganecologicalbalance(Tabashniket al.,2013).

InBurkinaFaso,cottonyieldlossesoftensurpass

30%infieldstreatedwithrecommendedinsecti-

cideapplications (Vitaleet al.,2016). Inaddition,

theemergenceof insecticide resistance triggers

intensified insecticide use that poses significant

health hazards for many farmers and laborers

andisthesourceofextensiveenvironmentalpol-

lution(Kooistraet al.,2006).

InmostSub-SaharanAfricancountries,yieldsof

500-700kg/haofseedcottonproducedunder

rainfed conditionsare typical for varietieswith

a yield potential close to 3,000 kg/ha.With ir-

rigation, cotton farmers can obtain yields of

4,000–5,000 kg/ha with high-yielding varieties

(ICAC,2017)(Figure2.2).Whiletheworldmarket

priceforcottonremainslow,smallholderfarm-

ers who rely on family labor to cultivate their

cotton and obtain yields of 600 kg/ha would

probablynotcovertheirinputcostiflaborwas

included in the calculation of net cost benefit.

Under thesecircumstances, it isnotsurprising

that national annual cotton production varies

greatly fromyear to year inmostSub-Saharan

African countries. And, althoughweatherplays

animportantpartinthevariation,thefarmgate

priceperkgofseedcottonintheprevioussea-

sonhasastronginfluenceonfarmers’decisions

to plant cotton the following season. The low

yieldsalso reflect the reality thatmanyAfrican

agricultural systems are far frombest practice

(Tripp,2009).

B

15

Bt cotton to limit crop losses due to insect attacksOne way to address insect damage in cotton would be to seek resistant varieties through conventional plant breeding, but insect-resistant crop varieties have usually been much more difficult to develop than those for disease resistance. Insects often have less specialized nutritional habits than the microorganisms that cause plant disease, and are able to attack various crops.

3

Source: Bruno Tinland, personal collection

A major breakthrough in plant breeding for

insect-resistanceincottonwasachievedonlywith

theuseofgeneticallymodified(GM)varietiesthat

produceaproteintoxictospecificinsects.Indeed,

thesoilbacteriumBacillus thuringiensis(usuallyab-

breviatedto‘Bt’)producesproteins(forinstance,

Cry1AcandCry2Ab) thatareonly toxic to some

moth and butterfly caterpillars and/or larvae of

beetles and mosquitoes. They are harmless to

other insects and animals, including humans.

Once ingested by, for example, the cotton boll-

worm,theCryproteinsbindtospecificreceptorsin

theliningofthecaterpillar’sgut,wheretheycreate

holesandquicklycausedeath(HöfteandWhiteley,

1989).Bttoxinshavebeenknownforalongtime

andarethebasisofanumberofcommercialin-

sectcontrolproductsthatareparticularlypopular

withorganicfarmers.Inthiscase,whenBtisused

asa sprayingagent, it alsokills sensitive insects

thatdonotfeedontheplant(formoreinforma-

tion,seeVIBFactsSeriesissue‘BtcottoninIndia’).

Moreover,theBttoxinhasashorthalf-lifewhen

placedunderfieldconditionsduetodegradation

byUVlightandotherenvironmentalfactors.Thus,

many typesof insect larvaemayescape control

by theseproducts if spray coverage is not opti-

mal,because theyarewashedoffwhenapplied

or because the insect is already feeding inside

the plant and is thus protected from spraying

(Aronson,1986).

WEST AFRICAN COTTON FARMERS IN A CHANGING WORLD“Under a clear November sky, a group of West African farmers takes a break from harvesting their cotton. The

men survey the crop and dare to hope that the harvest will be better than last year, when drought meant they

were barely able to repay their loans for the expensive inputs used to produce cotton. The women participate in

the harvest even though some of their own food crop fields still need attention and there are scores of tasks to be

done at home. They need a good harvest, because cotton offers one of the few possibilities to earn the cash that

is used to pay school fees and buy medicine and other essentials.

In addition to their concerns about the harvest and

the price they will receive, these farmers now find

themselves at the center of a worldwide controversy

about agricultural biotechnology. The news they hear

on the radio and in discussions with other farmers

is difficult to interpret, and the debates mostly take

place in distant locations. The farmers hear there is

a new type of cotton that resists some insects and

reduces the need to buy insecticides. Some people ar-

gue that this will help them save money and keep up

with other cotton-producing countries, while others

say that it will put them at the mercy of powerful foreign companies and untested technologies.” (Quoted from

‘Biotechnology and Agricultural Development. Transgenic Cotton, Rural Institutions and Resource-Poor Farmers’

by Robert Tripp, 2009)

17

Researchersreasonedthatifthegeneproducing

theBtproteincouldbe inserted intotheplant’s

DNA,theplantcouldproduceitsowntoxin,killing

onlytheinsectsthatfeedonthecropwithoutthe

needforexternalinsecticidespraying.Bytheear-

ly1980s,anumberofpublicandprivateresearch

entitieswereattemptingtoproduceGMtobacco

plants containing a Bt gene. At the same time,

therewasincreasedpressuretopatentthegenes

codingforvariousBttoxins.Earlysuccesseswere

registered atWashingtonUniversity in St. Louis

andatPlantGeneticSystems,aGhentUniversity

spin-offcompanyinBelgium.Althoughitwaspos-

sibletodemonstratethepresenceoftheBtgene

intheGMplants,itsinsecticidalperformancewas

verymodest.WhenthelaboratoriesofMonsanto

discovered in 1988 that the sequence of the

bacterial gene needed to be optimized to be

compatiblewith those of plant systems, the re-

sulting insecticidal activity increased significantly

(Charles,2001).

ThedevelopmentofBtcottoncultivarsissimilarto

conventionalcottonbreeding,withtheexception

ofthefirststepinvolvingtheinsertionofBtgenes.

The insertionofthegeneof interest isachieved

usingoneof twomethods:Agrobacterium-medi-

ated transformation and particle bombardment

(Potrykus et al., 1998). The biggest drawback to

these techniques is the required use of plant

cellorin vitrotissueculturestointroduceagene

intotheplant’sDNA.Scientistsoftenchoosethe

American Coker cotton cultivar as the recipient

plantbecause itcanrecoveraplant fromtissue

culture, which is not the case for most cotton

cultivars(Smithet al.,2004).Thenewgeneinthe

AmericanCokercultivarcanthenbetransferred

tolocalcultivarsofinterestthroughconventional

breeding.

Herbicidetolerancewasanothermajordevelop-

mentofGMcotton technology,with thegoalof

producingcottonvarietiesabletotolerateappli-

Figure 3.1: In vitro tissue cultures required to introduce a gene of interest into the plant’s DNA (A) tissue culture; (B) growth chamber.

A

cationsofparticularherbicides.Herbicideshave

beenusedformanyyearsinindustrializedcoun-

tries,andincreasinglyindevelopingcountries,to

controlweeds(Naylor,1994).Manyherbicidesare

so-called‘broadspectrum’,killingawiderangeof

plants, and thus must be used before planting

orbyprotectingthestandingcropfromcontact.

Herbicide-tolerantGMcottonfieldscanbetreat-

ed with herbicide after the crop has emerged.

The development of herbicide-tolerant varieties

hasmadeconservationtillageevenmorefeasible

(Thompsonet al.,2007).Conservationtillageisa

cropmanagementsystemthathelpspreventero-

sionandreducesthenumberoftimesmachines

must enter thefield.As a result, it lowers costs

and reduces the riskof soil compaction.Cotton

fieldsareincreasinglyplantedundersometypeof

conservationtillage,includingtheuseofplanting

patternssuchas ‘skiprow’ thatcanhelpreduce

down-the-rowinputcostsandimproveweedcon-

trol through thewiderdistancebetween cotton

rows.Cottonplantsnexttotheskipwillcompen-

satebygrowinglargerandproducingmorefruit

(Thompsonet al.,2007).

In addition, there are GM cotton varieties with

both insect resistance and herbicide tolerance,

oftenreferredtoas‘stacked’Bt/herbicide-tolerant

varieties. In industrialized agriculture, herbicide

toleranceisthemostimportantGMtraitcurrently

inuse.Wherestackedcottonvarietiesareavail-

able, theyareoftenmorewidelyused than just

Bt or herbicide-tolerant varieties. For instance,

100%,97%,95%,80%and42%ofcottonplant-

ed in 2016 was stacked cotton in South Africa,

Australia, Mexico, USA and Brazil, respectively

(ISAAA,2016).

B

19

Development of Bt commercial cotton cultivars and hybridsAfterBtcottonplantshavebeenrecoveredfrom

tissueculture,arigorousselectionprocessisun-

dertaken to identifyplantswithgoodagronomic

characteristicsandthehighestandmostconsis-

tentlevelsofBtgeneexpression.Themostsuitable

Bt plants are typically allowed to self-fertilize for

a few generations to ensure that inheritance of

theBtgene ispredictableand itsexpressionre-

mains stable (Skinner et al., 2004). During this

stage,plantbreedersmayalsoselect individuals

with particularly good agronomic characteristics.

Anothergoaloftheself-fertilizationprocessisto

produce cotton plants that are homozygous for

thegene(i.e.plantsthathaveacopyofthegene

atthesamelocusoneachhomologouschromo-

some).Homozygousplantsaresometimescalled

‘true-breeding’. The end result of this selection

processisatrue-breedingBtcultivar.

This true-breeding Bt cotton cultivar is rarely

commerciallyusefulbecauseofitsCokergenetic

background (Smith et al., 2004). As described

above,theAmericanCokercottoncultivarissuit-

ableforGMmethods,butitdoesnothavegood

agronomiccharacteristics.Todevelopacommer-

cial Bt cotton cultivar and eliminate the Coker

geneticbackground,aseriesofbackcrossesare

conducted (formore information, see VIB Facts

Seriesissue‘Fromplanttocrop:thepast,present

andfutureofplantbreeding’).Thisbeginswhen

theBtCoker line is crossedwithanestablished

commercialcultivar(Figure3.2).Theinitialproge-

nyexpressestheBtgene,butstillcontainshalfof

theCokergenes thatmayconferdifferentchar-

acteristics than theonesof thedesiredcultivar.

Topurify thenew cultivar, this progeny is back-

crossedwiththecommercialparentcultivar.After

five generations of backcrossing, about 98% of

theundesiredCokergenesarereplacedbygenes

of the commercial cultivar. Thus, backcrossing

for 5-10 generations eliminates nearly all Coker

genes and produces a Bt cotton cultivar that is

similartotheoriginalcommercialcultivar,except

forthepresenceoftheBtgene(DuckandEvola,

1997).Thenextstepistoself-fertilizeplantsthat

bear the newly inserted gene and retain prog-

eny homozygous for the Bt gene. This yields a

true-breedingcommercialBtcottoncultivar.

Figure 3.2: An illustration of a backcrossing scheme aiming

to transfer the Bt gene (shown as an orange dot) from one

plant (e.g. the Bt Coker cotton cultivar, shown as green dots) to

another plant (e.g. an established commercial cotton cultivar,

shown as purple dots), while eliminating in the crossing product

as many of the undesirable characteristics as possible.

Cotton in Africa

ThedevelopmentofBt cottonhybrids canbea

valuableoption for growers inpartbecausehy-

bridsareoftenmorevigorousthaneitheroftheir

parental cultivarsandcan increasecottonyield.

Consequently,theycanbeplantedatlowerden-

sities,saving laborandcosts (Donget al.,2004).

Theprocesstodevelopsuchhybridsrequiresan

additionalstep.Thetrue-breedingcommercialBt

cultivar(asdescribedabove)iscrossedwithadif-

ferentcultivartoproducehybridseed(formore

informationonhybrids,seeVIBFactsSeriesissue

‘Fromplanttocrop:thepast,presentandfuture

ofplantbreeding’).Untilrecently,commercialpro-

ductionofcottonhybridswasdifficulttoachieve.

Because cotton plants mainly self-fertilize,

hybrid seed production relies on careful hand

pollination,aprocess that is labor-intensiveand

potentially uneconomical. Advances in breeding

(e.g. inducingmalesterilityandfertility)haveim-

provedtheefficiencyofthisprocess(Zhanget al.,

2000). In 2016, India was the only country that

grewsignificantamountsofhybridBtcotton,with

approximately 95% of cotton farmers planting

andbenefitingsignificantlyfromBtcottonhybrids

(ISAAA,2016).

Production of Bt cotton cultivars and hybrids and their seed provisionTrue-breedingcommercialBtcottoncultivarsare

grown throughout the world (see table 3.1 be-

low).Seedcompaniesmaintainpurelinesoftheir

varieties and grow them in carefully monitored

fields to obtain subsequent generations of

high-qualitycommercialseed.Mostcottonfarm-

erschoosetobuyfreshcertifiedseedeachyear

(Figure3.3).Thosewhochoosetosavetheseed

of true-breeding commercial Bt cultivars from

theirownfieldsareusuallyabletopreservethe

qualitiesofthecottonvariety.However,cross-pol-

linationwithothercottonvarietiesmaycausean

overall decline in seed quality if it occurs (Heu-

bergeret al.,2008).

In the case of the production of hybrid cotton

seed (GM and conventional), seed companies

mustmaintainbothparentalcultivarsandcross

thecultivarseachyear.Iftheseedsfromahybrid

cottonaresavedafterharvest, theyareunlikely

toperformaswellasthehybridsthemselves,be-

causetheprogenyofthehybridswillexhibithigh

geneticandphenotypic variation.HybridBt cot-

tonplantscontainonecopyoftheBtgene,which

implies that approximately 25% of saved seeds

from hybrids will not carry the Bt gene. Thus,

savedseedfromBthybridcottonwillnotprovide

thesamelevelofprotectionagainstinsectsasthe

previous year’s hybrids (Kranthi et al., 2005). To

avoid theseproblems,hybridseedsareadvised

tobeobtained fromseed companieseach sea-

sontoensurehighperformance.

Cotton seed is somewhatmore difficult to save

from season to season than the seed of most

fieldcrops,asmechanicalseparationoftheseed

fromthefiberisrequired.Thismeansthatfarm-

ersmusteither reserveandbuybackaportion

oftheirseedfromtheginneryorhaveaccessto

small,hand-turnedginsthatallowhomeprocess-

ingoftheseed.Suchseed,evenifitistheproduct

ofasingleharvest,maybequitevariable,because

cottonisanindeterminatecropinwhichseedde-

velopmentisnotsynchronous,butratherspread

overaperiodoftime.Today,formalseedproduc-

tionbyprivateandpublicentitiesaccountsforthe

majorityoffarmers’cottonseed.InmuchofWest

Africa,seedisprovidedbyparastatalenterprises

and farmers have few incentives to save seed

21

Figure 3.3: (A) Cotton seeds on the plant; (B) bales of cotton waiting to be exported in Burkina Faso (source: Bruno Tinland, personal

collection); (C) cotton seeds after ginning.

(Figure 3.3). India’s strong public seed system

hasbeen largely replacedbyprivate companies

driven by the opportunity to sell hybrid seed,

which farmers have difficulty saving because of

the lower performance of hybrid progeny. Until

theadventofhybrids,however,therewasconsid-

erableseedsaving.Morethanhalfofthecotton

sowninIndia’sPunjabwasfromfarm-savedseed

asrecentlyasthe1990s(Sidhu,1999).

Commercialization of Bt cotton in Africa is spreading more slowly than in other cotton growing regionsBy1990,Monsantohadthefirstexperimentalva-

rietiesofBtcottonavailablefortesting.Thefirst

commercial plantingsofBt cotton tookplace in

1996intheUSA,MexicoandAustralia.Monsanto’s

firstBtcottonvarietiesweresoldwiththetrade-

markBollgardII®(containingtwoBtgenes,Cry1Ac

andCry2Ab).Sincethattime,severalothertypes

ofBt insect-resistant cottonhavebecome com-

C

A

B

Cotton in Africa

merciallyavailable(formoreinformation,seeVIB

FactsSeriesissue‘BtcottoninIndia’).

Based on the latest FAOSTAT data of 2014

(FAOSTAT,2017), cottonwasplantedon35mil-

lionhectaresglobally,64%(22.3millionhectares)

of which was Bt cotton. A total of 14 countries

grewBtcottonin2016ofthe26GMcropcoun-

tries worldwide (Table 3.1). India accounts for

the largest area of Bt cotton in the world, with

more than 10million hectares planted in 2016

(ISAAA,2016).

By 2016, at least four African countries have at

somepoint inthepastplacedaGMcrop inthe

market – Burkina Faso, Egypt, South Africa and

Sudan. However, due to various political and

technical setbacks,onlySouthAfricaandSudan

planted biotech crops, including Bt cotton, in

2016. InEgypt,abanonBtmaizewas imposed

becauseofclaimsonsafetyproblems,whilethe

government of Burkina Faso temporarily sus-

pended the growing of Bt cotton to address

a concern about fiber length observed in the

varieties farmershavegrownover the lasteight

years(seepage25).

SouthAfricawas thefirstcountryon theAfrican

continenttoadoptGMcropsforcommercialpro-

duction(Gouseet al.,2003).Thefirstfieldtrialsof

GMcropswereinitiatedin1989,andSouthAfrica

releaseditsfirstcommercialGMcropsin1998with

insect-resistantBollgardIIcotton.Thecommercial

release of Bt cotton was made possible by the

GeneticModifiedOrganismActof1997,and,inthe

1997-1998season, a few farmers in theMakha-

thini Flats grew the variety as a trial. Adoption

rapidly expanded following the first commercial

release,andby2001-2002,itwasestimatedthat

approximately 90% of farmers were growing Bt

cotton.In2016,9,000hectaresofBtcottonwere

planted,a25%decreasecomparedto2015due

todroughtandlowglobalcottonprice.Allcotton

grown in South Africa isGMwith 100% stacked

Bt/herbicide-toleranttraits(ISAAA,2016).

Cotton is one of themost important Sudanese

cropsandwasthemainforeignexchangeearner

beforethedevelopmentofanationaloilindustry.

SudanapproveditsfirstGMcrop-insectresistant

Btcotton-forcommercialplantingin2012,with

a single variety under the trade name Seeni 1.

SeeniisaChineseBtcottongenotypecarryingthe

Cry1AgenefromBacillus thuringiensis.Continuous

researchover the last five years resulted in ap-

provaloftwonewinsect-resistantcottonvarieties

in2015,graduallyincreasingthehectaragefrom

an initial launch of 20,000 hectares in 2012 to

120,600hectares in2016. In just five years, the

country has recorded a 98% adoption rate of

Bt cotton by Sudanese farmers. The two new

insect resistant cotton varieties from India,Hin-

di 1 released for irrigated regions and Hindi 2

forrainfedareas,haverecordedyieldsoftwoto

three times those of local non-Bt varieties and

significantly higher than the released Bt variety

Seeni1(ISAAA,2016).Table 3.1: Countries growing Bt cotton in 2016

(source: ISAAA, 2016).

Country Area (million hectares)

India 10.8

USA 3.7

Pakistan 2.9

China 2.8

Brazil 0.790

Australia 0.405

Argentina 0.380

Myanmar 0.325

Sudan 0.120

Mexico 0.097

Paraguay 0.010

Colombia 0.009

SouthAfrica 0.009

CostaRica Only210hectares

23

AtthebeginningofthethirddecadeofGMcrop

commercialization, the drive for research and

regulatory support for GM crops in Africa re-

mainedfocusedonfoodsecurity.Modernization

oftheagriculturalsectortomakeitmoreefficient,

competitiveandadaptedtoclimatechangealso

dominateddiscussionsatthepolicylevel.Onthe

other hand, many African governments priori-

tizedBt cottonasa strategically important crop

to revive the once-vibrant textile industry and

tapemploymentopportunitiesforyoungpeople

withinthecottonvaluechain(Figure3.4)*.

Significant milestones in the GM research and

biosafetypolicylandscapewereachievedin2016

inAfrica(ISAAA,2016).Atotalof8countrieseither

planted, actively evaluated field trials or moved

towardsgrantapprovals for thegeneral release

ofBtcotton(Table3.2).Kenya,MalawiandNigeria

movedfromconductingexperimentalresearchor

confinedfieldtrialstograntingapprovalsforen-

vironmentalrelease(opencultivation).Thiscould

lead to commercial planting in the next one or

twoyears,aftervarietalandnationalperformance

trials are completed in different agro-ecological

zoneswherethecropswillbegrown.Supportive

policiesareessentialtomakethishappen.Ethiopia

andSwaziland conductedmulti-location trials in

preparation for general release approvals that

willtakeplaceaftersuitabilityisascertainedalong

variousattributessuchasyield,levelofresistance

andlintquality,amongotherparameters.Agen-

eralreleaseapplicationforBtcottoninCameroon

is under way (ISAAA, 2016). Table 3.2 also cap-

tures thedevelopmentswithinSouthAfricaand

Sudan, which sustained commercial planting of

GMcottonin2016.

Figure 3.4: Young Burkinabe farmers playing in cotton lint

(source: Bruno Tinland, personal collection).

* http://farmbizafrica.com/profit-boosters/

african-governments-must-prioritize-agriculture-to-drive-

inclusive-economic-growth-and-development.

Cotton in Africa

Table 3.2: Ongoing GM cotton research activities in Africa by December 2016 (source: ISAAA, 2016).

Country Trait Institutions/companies involved Stage as of December 2016

Cameroon Insect-resistance/ herbicide-tolerance

BayerCropScienceandSODECOTON(LaSociétédeDéveloppementduCoton)

Applicationforenvironmentalreleaseinprocess

Ethiopia Insect-resistance EthiopianInstituteof AgriculturalResearch(EIAR);seedsobtainedfromJKAgriGenetics-India

Multi-locationtrialsin6sites

Kenya Insect-resistance KenyaAgriculturalandLivestockResearchOrganization(KALRO),Monsanto

Conditionalapprovalforenvironmentalrelease;toconductnationalperformancetrials

Malawi Insect-resistance LilongweUniversityofAgricultureandNaturalResources(LUANAR),DepartmentofAgriculturalResearchServices(DARS),Monsanto,Quton seedcompany

Generalreleaseapproved;varietyregistrationtrials underwaytobeplantedin 9sites

Nigeria Insect-resistance Monsanto Approvedforcommercialrelease

SouthAfrica Insect-resistance/ herbicide-tolerance

BayerCropScience Trialpermitgranted

Sudan Insect-resistance:2Indianhybrids1ChinesevarietySCRC37

BiotechnologyandBiosafetyResearchCenter;China-aidAgricultu-ralTechnologyDemonstrationCenter,theSudancottoncompanyElfaw

Multi-locationtrialscompletedfor3additionalBthybrid varieties;approvedfor commercialplanting

Swaziland Insect-resistance SwazilandCottonBoard, JKAgri-Genetics

Confinedfieldtrials approvalgranted

What happened with Bt cotton in Burkina Faso?BurkinaFasohasemergedasoneoftheleading

adopters of agricultural biotechnology in Sub-

Saharan Africa. The introduction of Bt cotton in

Burkina Faso corresponded with a severe crisis

in the cotton sector. In the mid-2000s, a weak

US dollar, high input costs, considerable cotton

subsidies in the developed world and a declin-

ingworld cottonmarketprice all contributed to

severeproblems inWestAfricancottonsectors.

The adoption of Bt cotton in Burkina Faso took

alsoplaceinthecontextofaseverelyunderfund-

ed public agricultural research sector. Burkina

Faso’s main research center INERA (Institut de

l’EnvironnementetRecherchesAgricoles)mostly

reliesonfundsfromthecottonsectortoperform

cottonresearchactivitiesincludingcultivarselec-

tionandseedpropagation.Inaddition,insecticide

resistancehademergedinBurkinaFasoand,asa

consequence,theefficiencyofconventionalpest

controlmeasureshasdecreased.It iswithinthis

context that Bt cotton represented a potential

benefitasanewpestcontroloptionforthestrug-

glingBurkinabecottonsector.

The Burkinabe government developed a legal

framework to regulate the field testing and

commercialization of GM crops. After several

years of field trials (2003-2007), the National

Biosafety Agency authorized Bt cotton for seed

production and commercialization in 2008

(Vitale and Greenplate, 2014). A key feature of

this decision allowed INERA to cross

25

Monsanto’sBollgardII®Btcottonwithtwolocally

used cultivars (Traoré et al., 2008). This result-

ed in thedomesticproductionofBtcottonseed

and improved the adaptability of Bt cotton to

local growing conditions. By 2014, the adoption

ofBtcottonhadalreadyapproached80%.2015

was the eighth year that farmers could grow

Bt cotton in Burkina Faso. A total of 350,000

hectares of a total cotton planting area of

700,000hectares,or50%ofthetotal landavail-

able,wereplantedwithBtcotton.Thisrepresents

a 23.8% drop in Bt cotton adoption from the

73.8%in2014(James,2015).Theanxietycreated

bytwocoupsinaspanofoneyear,subsequent

government transitions,alongwithhighBtseed

pricesandgovernanceissuesincludingcorruption

and late paymentsmay have persuaded cotton

producers to reduce the size of their Bt cotton

fields (Dowd-Uribe, 2014). Furthermore, a con-

cernraisedamongseedproducers,ginnersand

Burkinabe farmers over cotton fiber length led

tosomeuncertainties.Specifically,someginners

Cotton in Africa

reporteddeclinesinbothfiberlengthandginning

ratios(percentageoffiberperunitweightofcot-

tondelivered to thegin) inBt cottoncompared

to somehistorical conventional cotton varieties.

This decline in staple length has undermined

the reputationofBurkinabehighquality cotton,

andcutintoitsvalueontheinternationalmarket

(Figure3.5).

Whencoupledwiththedeclineinoveralllintdue

to the lower ginning ratio, the inferior quality

characteristics of the Bt cultivars have compro-

misedtheeconomicpositionofBurkinabecotton

companies(Dowd-UribeandSchnurr,2016).Asa

consequence,BurkinaFasohasputatemporary

halt on Bt cotton in 2016 to address the short

fiber length issue. Some options are potentially

promising and, if confirmed, might be available

commercially by 2018 (James, 2015). The Inter-

ProfessionalCottonAssociationofBurkina(AICB)

andthegovernmentreaffirmedtheircommitment

tobiotechnologyandgaveanassurancethatthe

concernwasnotwiththetechnologyBollgardII®,

but with the short fiber length resulting from

insufficientbackcrossesrealizedtointrogressthe

Btgeneintolocalvarieties*.Breedersandother

stakeholders are currently working towards

addressingthistechnicalissuewithintheshortest

timepossibletorestoretheBtcottonprogramin

thecountry.

A significant lesson learned fromBurkina Faso’s

caseistheimportantrolethattechnologydevel-

opers and breeders must play in incorporating

traits and qualities well-adapted to local condi-

tionsinordertomeetfarmerandmarketneeds.

ThefailedbreedingprograminBurkinaFasomay

alsocall intoquestion thepotential for combin-

ingGMtechnologyandlocalcottoncultivars,and

forinvestinginsuchlongresearchanddevelop-

ment efforts to produce new technologies that

offerdesiredperformanceacrossmultiplecriteria

(Dowd-UribeandSchnurr,2016).

Figure 3.5: Cotton lint in Burkina Faso

(source: Bruno Tinland, personal collection).

* http://www.biotechburkina.org/aicb-restates-support-to-gm-technology-but-acknowledges-that-farmers-will-feel-heat-of-temporary-bt-

cotton-suspension/ 27

Lessons learned from experiences in early-adopting Bt cotton countries worldwideMany articles have underlined the success of Bt cotton at reducing insecticide use, boosting yields, and increasing profits for millions of smallholder producers. Examples of this success originate in the two largest Bt cotton growing areas of India and China, and in Africa, with South Africa and Burkina Faso (Krishna and Qaim, 2012; Pray et al., 2002; Morse et al., 2004; Pertry et al., 2016; Vitale et al., 2016). In these countries, the cotton sector is largely dominated by smallholder farmers, who benefit from Bt technology adoption in the form of higher incomes and low exposure to health hazards associated with insecticide sprays. Other studies, however, questioned this unmitigated success. They demonstrated that though in many cases smallholder producers in the global south benefit from Bt cotton adoption, outcomes can be variable and success depends on a mix of institutional, socio-economic, and agro-ecological factors (Glover, 2010; Dowd-Uribe, 2014).

4

Source: Bruno Tinland, personal collection

29

Comparing the results (yields, income) of those

farmers who use the Bt technology and those

who don’t is not straightforward. Agricultural

seasonsarecharacterizedbygreatvariability (in

rainfall,insectpopulations,etc.).Thus,theresults

fromanyoneyearmaynotberepresentative.In

addition,thefarmerswhoarethefirsttoadopta

newtechnologymayhaveotherpracticesorre-

sourcesthatsetthemapartfromtheirneighbors,

making a side-by-side comparison problematic.

Moreover,theadoptionofanewtechnologymay

be so rapid that there are few farmers left to

serveasacontrolgroup.Anotherfactorcaneven

be that farmers cultivating conventional crops

mightbenefitfromtheirneighborscultivatingGM

crops, as this might create a lower insect

pressureintheregion.Finally,atechnologysuchas

Bt insect-resistance is not so much yield-

enhancingasyield-protecting.Itsefficacywillde-

pendon the levelofpestattackand theuseof

otherpestcontrolpractices(Liuet al.,2015).Allof

thesefactorscanbeaddressedtosomeextentby

carefulsurveyandstatisticalmethods.Neverthe-

less,onemustbeawarethatthistypeofimpact

assessmentisanimperfectprocessandassump-

tionsshouldbetakenwithcare.

Changes in yield and insecticide useThemajorityoftheaforementionedarticlesindi-

categainsfromtheuseofBtcottonintermsof

reductionininsecticideuseand/oryieldincrease

and/orlowerproductioncosts.Theglobalfarmer

incomegainsfromtheuseofBtcottonduringthe

20-yearperiodfrom1996to2015havebeenesti-

matedatUSD52billion(ISAAA,2016).Thesegains

mainly resulted from increased yields thanks to

reduced crop damage, especially in developing

countries, but also from decreased input costs,

mostlyindevelopedcountries(Pertryet al.,2016).

Since1996, theuseof insecticideandherbicide

on theglobalGMcropareahas fallenby618.7

millionkgofactiveingredient(an8.1%reduction)

-mostofitduetolowerinsecticideuse-relative

to the amount expected if this crop area had

been planted with conventional crops (Brookes

and Barfoot, 2017). The environmental impact

associatedwithinsecticideandherbicideuseon

thesecropswasreducedby18.6%(asmeasured

bytheEnvironmental ImpactQuotient) (Brookes

andBarfoot,2017)(Figure4.1).

A

B

Figure 4.1: (A) Bt cotton that displays no damage from insect

attacks (source: Karim Maredia, personal collection); (B) workers

spraying conventional cotton fields in Zimbabwe.

800000

700000

600000

500000

400000ha

adop

tion r

ate i

n per

cent

age

Total Bt cotton Adoption rate of Bt cotton

300000

200000

1000002

29

65

58

51

69 70

02008 2009 2010 2011 2012 2013 2014

80

70

60

50

40

30

20

10

0

Cotton in Africa

In Burkina Faso, the large-scale adoption of Bt

cottonhasresultedboth inapositiveeconomic

and environmental impacts when compared

to conventional farming practices. By 2014, the

adoption of Bt cotton was approaching 80%,

the level consideredbymany in theproduction

literature as the long-term upper limit of new

technologyadoption(Figure4.2).Usingdatafrom

sixyearsof farmsurveys (2009-2014),Btcotton

wasshowntohavesignificantlyandconsistently

outperformedconventionalcottonyieldsineach

ofthesixyearsofcommercialproduction(Figure

4.3).Btcottoncultivationrequiredapproximately

two-thirdslessinsecticidethanconventionalcot-

tonandreducedfarmlaborallocatedtospraying

(Vitaleet al.,2016).Areductionininsecticideuse

resulted from the reduced annual numbers of

spraysfrom6to2.Twospraysarerecommended

tocontrolsensitiveinsectsstilldamagingthecrop

(Pertryet al.,2016)(Figure4.4).

Figure 4.2: Hectarage (bars) and adoption rate of Bt cotton in Burkina Faso (in %: orange line) (source: adapted from Pertry et al., 2016).

2014

900

800

700

600

500

400

300

200

100

0

500000

450000

400000

350000

300000

1200

250000

1000

200000

800

ha

tons

of ac

tive i

ngre

dient

Yield

(kg/

ha)

Year

Total ha

BGII

Bt cotton (ha) Insecticide use (tons of active ingredient)

CV

150000

600

100000

400

50000

200

0

0

2008

2009

2009 2010

2010 2011 2012 2013 AVE2014

31

Figure 4.4: Use of insecticides (orange line) in Burkina Faso since the introduction of Bt cotton (source: adapted from Pertry et al., 2016).

Figure 4.3: Comparison of Bt cotton (BGII: Bollgard II) versus conventional cotton (CV) yields in Burkina Faso (source: adapted from Vitale

et al., 2016 and reproduced with permission).

Cotton in Africa

Ifwe lookatChinaandAustralia,which contain

someoftheworld’slargestBtcottongrowingar-

eas,theadoptionofBtcottonhasalsoresultedin

markedly lower insecticideuse,butonlymodest

yieldincreases(QaimanddeJanvry,2005;Pemsl,

2006).WhileChinesecottonyieldswerealready

among the world’s highest, it is also important

to note that the significant instance of insecti-

cide reduction due to Bt cotton responded in

fact to a situationwhere farmers’ excessiveuse

ofinsecticideshadcreatedatreadmillofgrowing

insect-resistance and ever-increasing depen-

dence on chemicals. Although the quantities of

insecticideusedonBtcottonarenowmuchlow-

erinChina,theyarestillamongthehighestinthe

worldandincludesubstantial insecticideusefor

bollwormlateintheseason(Qiao,2015).

IncountriessuchasSouthAfricawherebothlarge

andsmallholderfarmershaveaccesstoBtcotton,

therearedifferences in resourcesandmanage-

ment practices between the two groups which

mayleadtodifferentimpacts.Large-scalecotton

farmersreportedinsecticideandapplicationcost

savings and peace ofmind about bollworms as

themajorbenefitsofBtcotton.Smallholderfarm-

ersindicatedfinancialsavingsoninsecticidesand

yield increasesas themajorbenefitand reason

for adoption of Bt cotton (Gouse et al., 2003).

ItwasalsoshownthatSouthAfricansmallholders

generally underuse insecticide and that, in this

case, theBt technology ismore yield-increasing

thaninsecticide-reducing.Iffarmersdonothave

the knowledge or resources to apply effective

insecticides, the adoption of Bt cotton some-

times has as much (or more) of an impact on

yields as it has on insecticide use (Shankar and

Thirtle, 2005). In those regions of the world

where chemical insect control alternatives are

inadequate or not available, the yield impact of

a crop like Bt cotton will be highest (Qaim and

Zilberman,2003).

In Burkina Faso, farm sizewas not found to be

adeterrenttoBtcottonadoption.Astudybased

onsixyearsoffarmsurveydatashowedthat,on

arelativebasis,farmsofallsizesbenefittedequal-

ly from growing Bt cotton, though larger farms

werefoundtobemoreproductiveandgenerated

largerabsolutebenefitsfromBtcotton(Vitaleet

al.,2016).

Technology costs need to be considered and their impacts can be variableThebenefitsof yieldgainsandreductions in in-

secticide use are welcome, although the net

economic impact depends to a great extent

on the technology cost. In cases where this is

high, the financial gains are sometimes less

thanexpected.

Owners of a technology want to earn asmuch

as possible from their innovations in order to

recover their investments and allow further re-

searchandexpansion.Thenormalprice for the

seed is leviedwith a ‘technology fee’ as a sepa-

rate transaction that in effect licenses the gene

toeachfarmer(Charles,2001).Consequently, in

South Africa, for instance, Bt seeds are about

three times as expensive as non-Bt seeds. ABt

cotton grower in Burkina Faso experiences on

average a production cost quasi equivalent to

a conventional cotton grower (319 USD/ha to

312USD/ha,respectively)(Pertryet al.,2016).This

wascalculatedoveraperiodoffiveyears(2009-

2013)byconsideringthegrossincomebasedon

yield,thesalepriceofcotton,aswellastheaverage

Bt co

tton p

rofit

abilit

y ($/

ha)

33

input costs for seeds, fertilizers, insecticides,

herbicidesandlabor.Thisinsignificantdifference

inproduction costs isdue to the fact that even

though Bt cotton farmers have a relevant gain

in fewer insecticide treatments, they incurhigh-

er seedcosts.Nevertheless, farmersgrowingBt

cottonhavea65.1%highernetincomethancon-

ventionalcottongrowersthatcouldbeattributed

to the yield gains and concurrently increased

grossincome(Pertryet al.,2016)(Figure4.5).

Similarly,astudyinfourIndianstatesshowedthat

althoughBt cotton led to a significantdecrease

in insecticide expenditure, the high cost of the

Bt seeds eliminated this saving. Themajority of

farmersexperiencedanincreaseinnetrevenues

onlybecauseofhigheryieldsfromtheBtcotton

(Qaim et al., 2006). Many Indian cotton farm-

ers are used to paying relatively high prices for

cottonseedbecauseof theprevalenceofhybrid

seedsinthemarket.Sincehybridsareoftenmore

vigorous than either of their parental cultivars

and can increase cotton yield, the amount of

hybrid seedneeded toplant a hectare ismuch

lessthanwithnon-hybridseed.Thus,theimpact

oncostsofproduction isnotnearlyasgreatas

the seed price comparison would indicate. In

China,there isagreatdiversityofBttechnology

andcottonseedsources(publicandprivate,legal

andillegal)makingitparticularlydifficulttodeter-

mineseedprice. In2006, for instance,although

theofficialpriceforseedcontainingtheBtgene

wasapproximatelyUSD10/kg,Btcottonseedwas

availableforaslittleasUSD2/kg(Pemsl,2006).

Figure 4.5: Bt cotton profitability ($/ha) average over 2009-2013. BGII: Bollgard II (Bt cotton); Conv.: conventional cotton; Insect.: insecti-

cides; Fert.: fertilizers; Product.: production (source: adapted from Pertry et al., 2016).

BGII Conv.

600

500

400

300

200

100

0

56

10

SEEDS

22

65

INSECT.

119 116

FERT. +HERB.

122 121

LABOR COST

319 312

PRODUCT. COST

208

126

NET INCOME

527

438

GROSS INCOME

Cotton in Africa

Cotton is a demanding crop in terms of crop husbandry and pest managementCrop production is highly dependent on the

weatherand thecontributionofany technology

isalsoinfluencedbyclimaticconditions.Farmers

have always had to cope with this kind of vari-

ability,butbecauseBtcottoncultivationrequires

an investment in seeds and inputs before the

season begins without knowing how rainfall or

pestpopulationswillaffectproductionandcrop

management, it can representaserious risk for

farmers.Suchrisksaremoredifficulttobearfor

resource-poorfarmers.Thehighertheseedprice,

theriskierthedecisionforaresource-poorfarm-

er to grow Bt cotton. In high-income countries,

farmers commonly use insurance schemes to

protectthemselvesfromrisksandshocks.Unfor-

tunately,accesstothesetypesofformalfinancial

riskmanagementproductsisessentiallynonexis-

tent inmostruralareasofdevelopingcountries

(JensenandBarrett,2017).

Inaddition,likeanyotheragriculturaltechnology,

Bt cotton cultivation must be integrated within

existingfarmingsystemsandmayrequireadjust-

ments in cropmanagement. Relevant examples

describedbeloware the implicationsofvariable

pest pressure on the added value of Bt cotton,

the consequences of reduced insecticide use

onnon-targetpestsandtheneedforfurtherre-

search on the performance of Bt cotton under

marginalgrowingconditions.

Bt cotton benefits depend on pest abundance BecauseBtcottonisonlyeffectiveagainstsome

pests,andefficacyagainstapestmayrangefrom

modest tohigh, it follows that theperformance

andvalueofBtcottonwilldependontheabun-

dance of pests. For any pests, infestation levels

varyannuallyandaresubjecttoweathervariabil-

ity,whichcanthenaffecttheperformanceofthe

technology.Insomeofthecasesreportedinthe

literature(Qaimet al.,2006),theyearofapartic-

ular survey was unusually dry, pest infestations

andcottonyieldswere low,and theadvantages

offered by Bt cotton were relatively modest. In

othercases,rainfallwasatorabovenormal,yields

(andinsectpopulations)werehigh,andBtcotton

showedamoremarkedadvantage.Highrainfall

notonly increases insectpressurebutalso ren-

ders insecticides lesseffectivebywashing them

fromplants,givingBtcottonanextraadvantage

overconventionalcotton(Qaimet al.,2006).

Consequences of reduced insecticide use for non-target pestsReducedinsecticideuseonBtcottonthattargets

pestssuchasbollwormsmayattimesleadtothe

resurgenceofsecondaryinsectsearliercontrolled

bytheseinsecticides.Forexample,significantre-

duction intheuseofsynthetic insecticides inBt

cottonfavoredoutbreaksofmiridsandleafhop-

pers inChina (Wuet al.,2002;Menet al.,2005).

Before the introduction of Bt cotton, these had

been relativelyminor pests that had been con-

trolledbythesamesyntheticinsecticidessprayed

tocontrolcottonbollworm.Similarly,anincrease

inleafhoppersonBtcottonwasobservedinSouth

Africa,possiblyasaresultofreducedinsecticidal

spraysforbollworms(KirstenandGouse,2003).

Conversely, reduced insecticideuse inBtcotton

can have positive consequences for the control

ofpestswhosenaturalpredatorswere formerly

impactedbytheseinsecticides.

Based on data from 1990 to 2010 gathered at

36sitesinsixmajorcotton-growingprovincesof

northernChina,amarkedincreasewasshownin

abundanceofthreetypesofgeneralistarthropod

predators (ladybirds, lacewings and spiders) as-

sociatedwithwidespreadadoptionofBt cotton

andreducedinsecticidespraysinthiscrop(Luet

al.,2012).Anotherpositivespill-overeffectofre-

duced insecticideuse is thatwidespreaduseof

Bttechnologymaysuppressbollworminfestation

levelsregionally,suchthatnon-Btcottongrowers

mayalsobeabletoreducetheirinsecticideappli-

cations(KrishnaandQaim,2012).

Performance under marginal growing conditions WhenanewtechnologysuchasBtcottonisintro-

ducedindevelopingcountries,oneoftheprimary

concerns is itsperformanceunder themarginal

conditions and limited resourcesof smallholder

farmers. For example, the causes of the cotton

yield stagnation observed inWest Africa during

theperiodof1990–2005canbeexplainedby

several factors, such as agronomic constraints

from pest damage, poor soil fertility manage-

ment, lack of varietal development and climate

change. However, yields can also be negatively

impactedbytheexpansionofcottonproduction

into marginal lands where yields are inherently

lower (Vitale et al., 2011). Beginning in themid-

1980s,when the downward trend in yields first

became evident in West Africa, the national

cottoncompaniesrespondedbypushingbackthe

agricultural frontier, expanding into marginal

production areas. While the land expansion

strategy increased production in the short-run,

average yields continued to stagnate and even

fallintothe1990sand2000s(Vitaleet al.,2011).

Moreover, there is much controversy centering

aroundtheperformanceofBtcotton insomere-

gions of India. The problems reported with the

performance of Bt cotton seemed to result from

specificenvironmentalconditions,patternsofinsec-

ticideuse,lackofaffordablecreditandtheabsence

ofsocialsecurity.Inaddition,theexpansionofthe

cottonareaintoregionslesssuitableforcottonmay

alsohaveplayedarole(VIBFactSeries,2013).

Cotton in Africa

Marginallandsoftenhavepoorsoil(e.g.low-fer-

tility soil, high soil salinity) and are arid (low

wateravailability) (Figure4.6).Theyaretypically

characterized by low productivity and reduced

economic return. The crop harvest is doomed

to fail in theseareasregardlessofwhether the

crop is Bt cotton or conventional cotton. Yet,

there is increasing global interest using mar-

ginal land for bioenergy biomass production

aswellasfornon-foodcropssuchascottonin

the face of limited arable land resources. The

debate onmarginal land use is also a serious

topic in Africa associated with the trilemma of

landuseplanning:foodsecurity,bioenergy,and

environmentalconcerns(e.g.soilerosion,biodi-

versityloss)(Wendimu,2016).Theproductivityof

marginal landscanbeimprovedeitherthrough

technological improvement or through the use

ofnewcultivarsadaptede.g.todroughtandsa-

linity(Fitaet al.,2015).

Sustainability of the technology - how to manage the evolution of insect resistance to Bt cottonThe remarkable ability of insects to adapt to

insecticides supports the conclusion that the

developmentof resistancebypests is themain

threat to the continued success of Bt cotton

(Tabashniket al.,2013).Insectsmayalsodevelop

resistance to Bt toxins. Resistance to a toxin is

definedasageneticallybaseddecreaseinthefre-

quencyofindividualssusceptibletothetoxinina

population thathasbeenpreviouslyexposedto

thetoxin(Tabashnik,1994).Thepotentialforthis

typeofresistancewiththe increasingcultivation

ofBtcropsisoneofthemosthotlydebatedtopics

in agricultural biotechnology. Evenwhen appro-

priateBtcropsandcropmanagementpractices

havebeenidentifiedforlocalgrowingconditions,

resistancemanagementstrategiesareneededto

ensurethatBtcropsarecontrolledinsuchaway

Figure 4.6: Semi-desert region in South Africa.

37

Figure 4.7: Seed packages containing various cotton seed mixes for sale in China (source: Peng Wan, personal collection).

thatthedevelopmentofresistancetothetoxinis

kepttoaminimumtoenhancethesustainability

ofthetechnology(Tabashniket al.,2013).

In order to delay the emergence of insect re-

sistance to Bt toxins in cotton, various refuge

policies have been put into place. They require

theplantingof conventional (non-GM) cotton in

ornearBtcottonfields,ortheplantingofother

suitablecropstopromotethesurvivalofsuscep-

tibleinsects.Susceptibleinsectswillmatewiththe

rare resistant individuals surviving on Bt cotton

(Carrière et al., 2005). The use of plants pro-

ducing twodistinct toxinsmayalsodelay insect

resistance.Thisiscalledthepyramidstrategy.The

mostwidespreadcurrentexampleisBollgardII®,

whichproducesCry1AcandCry2Abgenes.

Refugemanagementiscomplex.However,anoth-

erpotentialconcernistheuseofseedmixtures.

In many countries, farmers save seed to plant

the next season. Cross-pollination with other

non-GMcottonvarietiesmaycauseanoverallde-

clineinseedqualityandBttechnologyefficiency

(Heubergeret al.,2008).Evenwhenfarmerspur-

chaseseedeveryyear, itmaynotbepurelyGM

or non-GM if the seed company’smanagement

of its production plots has been inadequate

(Figure4.7).Anadditionalsourceofconcernisthe

factthatwhenfarmersareprovidedwithnon-GM

seedtoplantasaseparaterefuge,theymayuse

ittofillgaps,thusunwittinglyplantinganinternal

refugewhen an external one is required. Alter-

native refuge schemes such as a ‘bag-in-a-bag’

(wherebytheGMseedismixedwithanappropri-

atepercentageofnon-GMseedbeforesale)have

beenproposed.However,thistypeofrefugehas

been shown to increase resistance toBt cotton

(Brévaultet al.,2015).

Figure 4.8: (A) Cotton integrated pest management training

in India (source: Regional Agricultural Economic Development

Centre); (B) Fairtrade cotton project in Burkina Faso.

Cotton in Africa

Beyond thebest strategies toavoid insect resis-

tance,thereareconcernsaboutthefeasibilityof

implementinganymandatorymanagementstrate-

giesinAfricanagriculturalsystems.Becausemuch

ofAfricanagricultureisinthehandsofsmallhold-

er farmers, the education and communication

requirements of any risk management strategy

are significant. Inmany countries, the owner or

licenseeofthetechnologyisalsoresponsiblefor

ensuring that the refugepolicy is enforced. This

is most often done via signed agreements with

seeddistributorsandfarmersthatobligatethem

to plant a refuge. This is also done by conduct-

ingon-farminspections.Incountrieswheresuch

measuresaredifficulttoenforce,themonitoring

(and compliance) is less in evidence. India has

dealtwith therefugechallengebyrequiring that

the seed companyprovides the farmerwith the

requisiteamountofnon-GMseedwhenpurchas-

ingBt cotton.However, thereare few resources

for monitoring compliance. In Argentina where

much of the Bt cotton planted is farmer-saved

(orobtained throughan informal sector), refuge

requirements are impossible to enforce in such

circumstances(QaimanddeJanvry,2003).

Another concern related to the sustainability of

GMcropsistheirpotentialeffectoncropagrobio-

diversity,whichcomprisesthediversityofspecies

usedinagriculturalproductionandvarietaldiver-

sitywithinthosespecies(Carpenter,2011).While

GM technologyallows the introductionofdesir-

ablegenesandtraitsintomanyexistingvarieties,

potentially making it easier to preserve varietal

diversity(Zilbermanet al.,2007),itisalsopossible

thatafewGMvarietiesmightdominateanation’s

croppingpatterns,therebyloweringtheresilience

provided by wider diversity and contributing to

agrobiodiversityerosion.Theissueofbiodiversity

lossisnotstrictlyassociatedwithGMcrops,but

canalsobetheconsequenceofmonoculture,re-

gardlessofwhetherthecropisBtorconventional

cotton.AstudyusingdatafromtheIndiancotton

sector and comparing levels of on-farm varietal

diversitybetweenfullBtadopters,partialadopt-

ers,andnon-adoptersdemonstratedthatcotton

varietaldiversity -withover96%adoptionofBt

cottonbyIndianfarmersin2016(ISAAA,2016)-is

atthesamelevelthanitwasbeforetheintroduc-

tionofthistechnology(Krishna,et al.,2016).

AnyyieldgainsfromtheadoptionofBtcottonwill

betheresultofimprovedpestmanagementand

will requirebestagronomicpractice toobtaina

yieldgainsufficienttocovertheadditional input

cost due to the Bt technology fee. A significant

amountofworkhasbeendedicatedtocropman-

agement techniques known as ‘integrated pest

management’ (IPM). IPM strategies for cotton

require careful assessmentof pest populations,

reduction(orinsomecases,elimination)ofinsec-

ticideuse,andtakeadvantageofbiologicalinsect

controlproducts inordertomaintainthe insect

A

B

population at levels below those causing eco-

nomicallyunacceptabledamageorloss(CropLife,

2014).State supportandcoordinationof IPM is

oftenessential,andappropriate incentivesmust

be inplace. Therehasbeen considerableprog-

ress in IPMmethods and techniques. They are

generally information-intensive, involve learning

and constant adjustment, andoften require co-

ordinationamong farmersand theparticipation

of variouspublic andprivate institutions (Figure

4.8).Asaresult,theyarenotlikelytospreadrap-

idly among smallholder farmers without strong

institutions tosupport thegenerationandman-

agementofknowledge(seeChapter5).

The impact of agricultural policies on cotton productionThe impact of an innovation such as Bt cotton should not be assessed only in terms of yields or production costs, but also with respect to the interactions with the institutions that govern farmers’ access to the technology. The viability of smallholder cotton production considerably depends on the provision of adequate technology, which in turn depends on a range of institutions. These institutions include public and private research organizations, input and credit markets, regulatory and intellectual property regimes, information provision and farmer organizations. The governance of these systems of information provision is as important for GM crops as it is for conventional technology. The ability to understand and control a technology is also an important factor in determining the impact of the Bt technology on resource-poor farmers.

5

Source: Bruno Tinland, personal collection

41

How the introduction of Bt technology has affected the organization of the input deliverySmallholdercottongrowersneedtoacquirethe

necessaryproductioninputstoboosttheirpro-

ductivity,regardlessofwhetherthecropisBtor

conventional cotton. However, they often lack

accesstoproductivity-enhancinginputssuchas

improved seed, fertilizers, water and informa-

tion. The credit needed to finance investment

intheseinputs isthemajorconstraint.Asare-

sult, smallholder farmers are unable to deliver

the volume and quality of product that com-

mercialbuyers–retailers,processorsandother

agri-businessfirms–require,which inturn lim-

its thedevelopmentofmarkets for agricultural

products.Thetypeofcreditavailableoftenplays

aroleindeterminingwhatinputsareavailableto

farmers.Theincentivesofthecreditsupplierare

amajorfactorindefiningthetypeoftechnology

employedbyfarmers.Theeffectivenessofinput

provision, and theextent towhich farmers are

able to understand and take advantage of the

inputsthatareonoffer,dependtosomeextent

onthenatureoftheinputindustriesthemselves.

Various strategies for input provision in cotton

cultivationhavebeendevelopedbyAfricangov-

ernmentsduring colonial times. For instance, in

Tanzania, cooperativeswereestablished topro-

videcredit,inputsandplowingservices,andthey

also had amarketingmonopoly. After indepen-

dence,thecooperativescameundergovernment

controlandgainedamonopolyonginningaswell

(Putterman, 1995). Suffering from mismanage-

mentand lackofcapital, theydeclined together

withthecountry’scottonproduction.

InseveralanglophoneAfricancountries, the lack

of alternative credit sources for cotton produc-

tion has led to the establishment of outgrower

schemes,whicharesystemsthatlinknetworksof

unorganized smallholder farmers with domestic

andinternationalbuyers.Alsoknownascontract

farming, these schemes provide benefits to all

playersalongthesupplychain.Buyerscanimprove

their control over crop supply, often at pre-

agreedprices, aswell as cropquality standards.

Farmers canaccessmore securemarkets,often

receivingtechnicalandfinancialsupportbyculti-

vating within outgrower schemes (TechnoServe

and IFAD, 2011). Ginning companies provide

inputsonloantofarmers,whoareexpectedtode-

livertheirharvestinreturn.Thesesystemsrequire

afinebalance.Excessivecompetitionencourages

farmers todefaulton their loansby selling their

cottontoarivalginnery.Controllingthemarketby

limitingthenumberofginneriesorprovidingter-

ritorial concessions canhelp reduce side-selling,

but heavy-handed coordination or monopolies

canresult in lowerpricespaidto farmers.Tobe

specific,side-sellingisasituationinwhichfarmers

aresellingtoafirmotherthanthefirmfromwhich

they received input credit. Typically, side-selling

isdonewiththepurposeofavoidingloanrepay-

ment(Tschirleyet al.,2009)(Figure5.1).

The samedilemmas existwithin the ‘filière’ sys-

temoffrancophoneWestAfricainwhichnational

parastatal cotton enterprises follow an ‘admin-

isteredmonopoly’model,witha legalmonopoly

on input provision and amonopsony (amarket

structure inwhichonlyonebuyer interactswith

manywould-besellersofaparticularproduct)on

thepurchaseof cotton from farmers. The com-

pany provides the production inputs as well as

technical advice to the farmers. After the grow-

ingseason, thecompanybuys theyieldatfixed

BT COTTON IN SOUTH AFRICA – A TECHNICAL TRIUMPH BUT AN INSTITUTIONAL FAILURESouth African smallholder farmers growing Bt cotton in the Makhathini Flats have received much attention in the

literature. At first, most of the peer-reviewed reports on Bt cotton adoption were favorable and have been used

to promote the technology in the rest of Africa (Ismael etal., 2002; Gouse etal., 2003; Kirsten and Gouse, 2003;

Morse etal., 2004). These reports of Makhathini’s success soon became emblematic of the potential that Bt cotton

could offer to poor smallholder farmers throughout Africa. When South African cotton smallholders began to ex-

perience economic losses, it was explained that this was due to a combination of consecutive seasons of drought

affecting Kwazulu-Natal and a change in the marketing arrangements (Gouse etal., 2005).

When Bt cotton was first introduced during the 1997-

1998 season, the Makhathini smallholders were

served by a single ginnery. This monopoly position

induced the ginning company to invest in a credit

scheme that allowed farmers to cover the input costs

and the technology fee for the Bt variety. With a sec-

ond ginnery licensed to operate in the same area,

competition for seed cotton became the priority, and

the number of loan defaulters rose dramatically. This

Cotton in Africa

pricesfromthefarmers,whocanpayofftheirin-

put credit, and takeson transportation,ginning,

andmarketing(TheriaultandSerra,2014).

Figure 5.1: Cotton lint collected and waiting to be delivered to the ginning company (A) in Burkina Faso (source: Bruno Tinland, personal

collection) and (B) in Malawi.

BA

led to the collapse of the input credit system. Without access to credit, the technology fee for the Bt variety became

unaffordable for many cotton farmers in Makhathini. By 2002-2003, the number of Bt cotton smallholders had

fallen by 82%, although promotion of irrigation by the new ginning company during this growing season has seen

numbers rise again with cotton lint yield increasing from less than 400 kg/ha to over 500 kg/ha (Fok etal., 2007).

The introduction of Bt cotton for smallholders in South Africa has thus been a ‘technical triumph but an institu-

tional failure’ (Gouse et al., 2005). Bt cotton was planted on 17,000 ha in 2006 and 10 years later, on 9,000 ha

only (ISAAA, 2016). Despite the declines in cultivated areas, adoption rates and production values, the optimism

that pervaded during the early years of Bt cotton growing in South Africa remains (Schnurr, 2012). Researchers

are unanimous that the environmental variability and institutional arrangements that hampered long-term

success in Makhathini do not detract from Bt cotton’s potential in other African environments. The lesson learned

from Makhathini is that farmers can benefit from technological innovation only if the correct infrastructures and

institutions are in place (Schnurr, 2012).

43

Institutional structures governing cotton production in Sub-Saharan AfricaTo strengthen the competitiveness of cotton

production,processing,andexportsinanincreas-

ingly demanding world market, governments of

mostcotton-producingcountriesinSub-Saharan

Africa began implementing sectoral reforms of

their cotton industriesasearlyas1990s,orare

considering doing so. These reforms generally

involvedisengagingthestateandfacilitatinglarg-

er involvementof theprivatesectorand farmer

organizations to ensure higher competition in

input and outputmarkets. The reforms also in-

volve improving productivity through research

and development and technology dissemina-

tion,andseekingvalueadditionthroughmarket

development and processing of cotton lint and

by-products(TheriaultandTschirley,2014).

AreportpublishedbytheWorldBank (Tschirley

et al., 2009) provides an assessment of reform

experience acrossnineof themain cottonpro-

ducing countries of Sub-Saharan Africa (Benin,

Burkina Faso, Cameroon, andMali inWest and

Central Africa; Mozambique, Tanzania, Uganda,

Zambia, and Zimbabwe in East and Southern

Africa). It highlights the range of institution-

al structures governing cotton production in

Sub-SaharanAfricaandshowshowthesestruc-

tures drive cotton sector performance. The

reportrecognizesthetradeoffsbetweencompe-

titionandcoordination.Competitionisimportant

toensureefficiencyandthe fairsharingofben-

efitsbetweenbuyersandsellers. Yet, toomuch

competitionwillmakeitdifficultforstakeholders

to engage in coordination, which is necessary

to provide important services such as quality

control, input credit, research and extension. A

well-functioningcottonsectorisonethatstrikesa

balancebetweenthesecompetingneeds,provid-

ingsufficientbenefitstoallstakeholderssothat

thesystemisabletomaintainitselfandgrow.

Based on the structure of the market for the

purchase of seed cotton and on the regulato-

Is competition allowed for the purchase of seed cotton?

Are there many buying firms, or few? Is there more than one cotton buyer?

Is each firm assigned an exclusive geographical area in

which to buy seed

Market-Based

Competite Systems

Local Monopoly

Concentrated Systems

Hybrid Systems

National Monopoly

Regulated

NoYes

Yes

No

Few No

Yes

Many

Cotton in Africa

ry framework inwhichfirmsoperate, thereport

identifies five typesof cotton sectors among the

nineAfricancotton-producingcountriesdescribed

above.Figure5.2laysoutthetypologybasedfirst

on a distinction betweenmarket-based and reg-

ulatedsectors,withthelatterreferringtosectors

in which free competition for seed cotton pur-

chase is not allowed. The second distinction is

based on the number of buyers of seed cotton:

manyorfewinthecaseofmarket-basedsystems,

andoneormorethanone inregulatedsystems.

Thesetwodistinctionsgeneratefoursectortypes:

(1) national monopolies, (2) local monopolies,

(3) concentrated market-based systems, and

(4) competitively structured systems. A fifth

category-hybridmarketstructures-includesthe

sectors that cannot be classified into oneof the

fourmaintypes.

InCameroonandMali,cottonsectorsaremanaged

byanationalmonopolyresponsibleforpurchasing

allcottonfromfarmersatfixedpan-regionalprices.

Figure 5.2: Decision tree for cotton sector typology (source: adapted from Tschirley et al., 2009; © World Bank, License: Creative Commons

Attribution license CC BY 3.0 IGO).

45

National monopolies are owned and operated

bypublicormixedcompanies.Cottonsectors in

BurkinaFasoandMozambiqueareorganizedinto

local monopolies, in which exclusive purchasing

rightsaregiventooneginningfirmwithinadelim-

itedgeographicalzone(Figure5.3).Concentrated

market structuresdefine cotton sectors, suchas

thoseofZambiaandZimbabweuntilat leastthe

early2000s,inwhichaverysmallnumberoffirms

(twointheseparticularexamples)dominatemar-

ket share but face free competition from other

firms and, potentially, from each other. Unlike

local monopolies, concentrated sectors have no

geographicalzoningthatdelimitsfirms’scopesof

operations. In competitively structured sectors,

suchasthoseinTanzaniasince1994andUganda

during the initial years of liberalization, a large

numberofbuyerscompetewithoutrestrictionto

purchaseseedcottonfromfarmers,withnosingle

setoffirmsdominating.Finally,hybridstructures

encompasscottonsectorsthatareeitherattempt-

ingtoliberalizeanationalmonopoly(e.g.Benin)or

tosolveunintendedconsequencesfromtheliber-

alization process (e.g.Uganda since shortly after

liberalization)(TheriaultandTschirley,2014).

Sector structure has a major influence on performanceCompetitive,market-basedsystemscanenhance

cottonproductionbyensuringrelativelyhighprices

tofarmers-adirectresultofintensecompetition

amongcompanies-butpoorlyperformon input

creditprovision,extensionandquality.Thisisbe-

causeofthedifficultyincoordinatingacrossmore

thanafewcompanies,whetherthiscoordination

istopreventside-sellingortoagreeondiscounts

tobepaidforpoor-qualityseedcotton.Asaresult,

competitive systems tend to generate low yields

andscorepoorlyonlintquality, limitingtheprice

advantage theyactuallypass to farmers.Monop-

olistic(nationalorlocal)andconcentratedsectors

canperformwellonpricespaidtofarmers.How-

ever,suchperformancedependsonthestrategic

Figure 5.3: Weighing of cotton harvested in Burkina Faso (source: Bruno Tinland, personal collection).

Cotton in Africa

prioritiesofdominantcompanies,ontheexistence

ofpoliticalinterference(ifany),andonthevoicesof

cotton farmers inpricenegotiations. Since2000,

concentrated sectors (Zambia and Zimbabwe)

haveperformedrelativelypoorlyregardingprices

to farmers, while nationalmonopolies have paid

unsustainably high (yet politically backed) prices

thathavebeenanimportantcontributortothese

sectors’fiscalcrises.Inthecaseofby-productval-

orization, the West and Central African sectors

(especiallythoseinMaliandBurkinaFaso)receive

lowpricesforcottonseed,anoutcomerelatedto

thehistoryofverticalintegrationwithinthesector.

In any case,by-product valorizationhas received

insufficientattentioninthestudyofcottonsectors

sofar.Itisanareawherevaluecouldbeaddedand

redistributedtofarmersthroughgreateropening

of thesectorandthroughcompetition.Monopo-

listicandconcentratedsectorsdobetteroninput

creditandservicestofarmers-andthusonyields

-aswellasonquality.Thesestructuresareable

todo thisdue to limited competition in theout-

putmarket,bettercoordinationamongfirms,and

consequently lower riskof side-selling compared

to competitive systems. Yet, theymay pass little,

ifany,ofthequalitypremiumontofarmers,and

theymaytendtochargehigher-than-marketrates

fortheinputstheyprovide.Consequently,returns

to farmers are similar in Zambia’s concentrated

system and Tanzania’s competitively structured

one(Tschirleyet al.,2009and2010).

Asaresult,nosinglemarketsectortypeseemsto

haveperformedsowellthatitcanbeusedasaref-

erenceforothercountries(TheriaultandTschirley,

2014). All sectors show productivity and perfor-

mancegaps.Thus, theygenerally lagwellbehind

the best performers in the world. Nevertheless,

sectortype(marketstructureandassociatedreg-

ulatory framework) does say a great deal about

the key challenges that will be most difficult for

a sector tomeet,andabout themostpromising

approachesfordealingwiththosechallenges.For

example,inputcredit,extension,andqualitywillbe

problems incompetitivesystems;prices to farm-

erswilltendtobelowinconcentratedsectors;and

companyefficiencywilltendtobepoorinmonop-

olies(Tschirleyet al.,2009,2010).

The implementationof reforms that are in line

with local realities faced by cotton farmers is

requiredtoenhanceefficiencyincottonproduc-

tion and revitalize Sub-Saharan African cotton

sectors, while inducing economic growth and

poverty alleviation (Theriault and Serra, 2014).

Improvement in the credit market requires

strongengagementfromgovernments,financial

institutionsandfarmers’associations.However,

access to credit is a necessary but insufficient

solutiontohighperformance.Toensureefficient

deliveryofinformationandadoptionoftechno-

logical innovationsaswellasbetteragricultural

practices, fostering farmer management skills

via functional extension services is as import-

ant as the establishment of adequate pricing

mechanismsthatensurereasonablepricesand

on-timepaymenttofarmers.

Farmers’ access to information on managing the Bt technologyTheway inwhich inputsaremadeavailablehas

implications for cotton farmers’ understanding

of,andcontrolover,Bttechnology.Whilefarmers

needtobeabletorecognizeandselectthemost

appropriatecottonvariety,theyalsoneedtoun-

derstandsomethingaboutthenatureofBtcotton

andhow it can contribute topestmanagement

strategies. Informationon variety characteristics

should be available from extension agents, the

47

commercial input systemor theadviceofother

farmers (Figure 5.4). In addition to information

provided through input markets, farmers also

require considerable information about crop

management. The stewardship in the case of

GMcrops requires the technologydeveloper to

provide downstream actors with the necessary

information and training for optimal use of the

material.Thiscanbedonedirectlybythetechnol-

ogyprovider,orthroughcontractors.

Farmers often rely on the experience of their

neighbors rather than on formal information

sources,andthesamepatternmayholdtrue in

thecaseofBtcotton.Along-termstudythatfol-

lows decision-making on Bt cotton in India has

foundthatalthoughfarmersrelyoninformation

fromeachotherintheirchoiceofcottonvarieties,

theprocessissubjecttofadsandrumorsrather

than being the product of careful experimenta-

tion.Itresultedinlargeshiftsinthepopularityof

individualvarieties fromoneseason to thenext

(Stone,2007).Thisstudyshowedthatfarmersare

oftenunabletodescribethebasiccharacteristics

(such asmaturity ormoisture requirements) of

thevarietiesthattheyhavepurchased.

Similarly,differences inpestcontrolpracticesare

notonlyduetodifferencesinfarmerincomeorna-

tionalinputmarkets.Forinstance,better-educated

Indian farmers make fewer and more selective

applications of insecticide (Qaim, 2003). In

Argentinaontheotherhand,wheremanyfarmers

under-investininputs,moreeducationiscorrelat-

edwithhigherinsecticideuse(QaimanddeJanvry,

2005). Farmer knowledge is certainly a crucial

element in the ability to use pesticides properly

or tobest take advantageof an innovation such

asBt cotton. It appears that someSouthAfrican

smallholders believed that Bt cotton provided

protectionagainstotherinsectsbesidesbollworms

and may have mistakenly changed their use of

insecticidesaccordingly(Bennettet al.,2006).

We have seen that farmers are sometimes bad-

lyservedby theguidancemadeavailable to them

through commercial seed and pesticide markets

and often have to contend with confusing or de-

ceptive information.On theotherhand, thereare

occasionallyinstanceswhenfarmersmaybeableto

deceivetheoutputmarket.Therefugemanagement

rules imposedonseedcompanies,whichtheyare

supposedtoenforce,maybeignoredbyfarmersei-

therbecauseoflackofunderstandingorinsufficient

follow-up.AstudyfoundthatevenmanyIndianseed

retailerswereunabletoexplaintherationaleforthe

smallpacketsof refugeseed theywereobliged to

sellalongwiththeBtcottonseed(Stone,2004).

Figure 5.4: (A) Cotton flowers and (B) a Burkinabe cotton plant variety (source: Bruno Tinland, personal collection).

BA

Conclusion6

Cotton in Africa

Throughout Sub-Saharan Africa, the cotton

sector faces major challenges regarding com-

petitivenessandsustainability,includingfinancial

crisesbroughtonby yearsof decliningproduc-

tivity throughout the sector. These crises are

associatedwithunfavorableexternalfactorssuch

as market distortions and credit default crises.

An innovation suchasaBt crop isnot simplya

technicalsolution.Itisaninterventionwithsocial,

economic and political consequences. Bt tech-

nologycancertainly contribute to the reduction

of insecticideuseand insectdamage, increased

yieldandhigherfarmerprofits,andtheconserva-

tionofbeneficialnaturalenemies.However,

moreattentionneedstobefocusedonthedevel-

opmentoflocalinstitutionsforfarmerstotakefull

advantageofthetechnology.Theselocal institu-

tionsincludetheinstitutionsthatsupportpublic

and private capacity for technology generation,

technology delivery through markets, extension

andregulations,andfarmercapacitiestodemand

services, participate inmarkets and understand

the technology they are using. Much broader

accessby smallholders to improved inputpack-

ages,suchasimprovedseedandfertilizerandto

thetechnicaladviceneededtousethemproperly,

isalsolargelyrequired.

Donors and governments that invest heavily in

thehopethatGMcropswillbringsignificantim-

provements to the livelihoods of resource-poor

farmers, without first paying attention to the

fundamental institutions that support broader

agriculturaldevelopmentand technologygener-

ation(withorwithoutGMcrops),areindangerof

misallocatingtheirresources.

ThecontributionofBttechnologytoinsectcontrol

inthecotton-producingcountriesofSub-Saharan

Africadependson:

1. Thedevelopmentofacommercialseedsec-

torandtheestablishmentandenforcement

of the basic regulations that provide an

enablingenvironmentforthatsector.

2. Thesupportofagriculturalresearch.

3. Awell-implementedcapacity-buildingstrate-

gy,sothatthepotentialbenefitsofBtcotton

areunderstoodandfarmersappreciatethat

makingaprofitwilldepend,evenmorethan

it did with conventional varieties, on the

implementation of best practices in

integratedpestmanagement.

4. An improved farmer access to credit for

input purchase and to technical advice.

Thepublic-sectorextensionservicedoesnot

have themanpower to deliver this and the

appropriate provider is the ginning compa-

ny. Yet, the conflict between cooperation

andcompetitionbetweenginningcompanies

needstobeaddressedinsomecountries.

5. The promotion of Bt cotton as a crop

protection technology and not primarily

asayield-enhancingone.

ReferencesAbhiyan,H.B.,Abhiyan,B.S.(2012).AdecadeofBTcottonin

MadhyaPradesh:Areport.Aronson,A.I.,Beckman,W.,Dunn,P.(1986).

Bacillus thuringiensisandrelatedinsectpathogens.Microbiology Review50:1-24.

Baffes,J.(2011).CottonSubsidies,theWTO,andthe‘CottonProblem’.The World Economy:1534-1556.

Bennett,R.,Morse,S.,Ismael,Y.(2006).TheeconomicimpactofgeneticallymodifiedcottononSouthAfricansmallholders:yield,profitandhealtheffects. Journal of Development Studies42:662-677.

Brévault,T.,Tabashnik,B.E.,Carrière,Y.(2015).AseedmixtureincreasesdominanceofresistancetoBtcottoninHelicoverpa zea. Scientific Reports5,9807; DOI:10.1038/srep09807.

Brookes,G.,Barfoot,P.(2017).Environmentalimpactsofgeneticallymodified(GM)cropuse1996-2015: Impactsonpesticideuseandcarbonemissions. GM Crops and Food8:117-147.

Carpenter,J.(2011).ImpactofGMcropsonbiodiversity. GM Crops2(1):7-23.

Carrière,Y.,Ellers-Kirk,C.,Kumar,K.,Heuberger,S.,Whitlow,M.,Antilla,L.,Dennehy,T.J.,Tabashnik,B.E.(2005).Long-termevaluationofcompliancewithrefugerequirementsforBtcotton.Pest Management Science 61:327-330.

Charles,D.(2001).Lords of the Harvest. Cambridge,MA:Perseus.

CropLifeInternational(2014).IntegratedPestManagement.https://croplife.org/crop-protection/stewardship/integrated-pest-management/

Dong,H.,Li,W.,Li,Z.,Tang,W.,Zhang,D.(2004).DevelopmentofhybridBtcottoninChina:asuccessfulintegrationoftransgenictechnologyandconventionaltechniques.Current Science 86:778-782.

Dowd-Uribe,B.(2014).Engineeringyieldsandinequality?Howinstitutionsandagro-ecologyshapeBtcottonoutcomesinBurkinaFaso.Geoforum53:161-171.

Dowd-Uribe,B.,Schnurr,M.A.(2016).Briefing:BurkinaFaso’sreversalongeneticallymodifiedcottonandtheimplicationsforAfrica.African Affairs115(458):161-172.

Duck,N.,Evola,S.(1997).Useoftransgenestoincreasehostplantresistancetoinsects:OpportunitiesandChallenges.InCarozzi,N.,Koziel,M.(eds) Advances in Insect Control: The Role of Transgenic Plants,Bristol,PA:Taylor&FrancisLtd,pp.1-20.

FAO(2012).Cropyieldresponsetowater.FAO irrigation and drainage paper No. 66. Rome.

FAOandICAC(2015).MeasuringSustainabilityinCottonFarmingSystems:TowardsaGuidanceFramework.Rome,2015.

FAOSTAT(2017).Availableathttp://www.fao.org/faostat/en/#data.

Fita,A.,Rodríguez-Burruezo,A.,Boscaiu,M.,Prohens,J.,Vicente,O.,(2015).Breedinganddomesticatingcropsadaptedtodroughtandsalinity:anewparadigmforincreasingfoodproduction. Frontiers in Plant Science6(978):1-14.

Fok,M.,Hofs,J.L.,Gouse,M.,Kirsten,J.F.(2007).ContextualappraisalofGMcottondiffusioninSouthAfrica. Life Sciences International Journal1:468-82.

Gadelha,I.C.N.,Fonseca,N.B.S.,Oloris,S.C.S.,Melo,M.M.,Soto-Blanco,B.(2014).Gossypoltoxicityfromcottonseedproducts.The Scientific World Journal:11p.

Glover,D.(2010).ExploringtheresilienceofBtcotton’spro-poorsuccessstory.Development and Change41(6):955-981.

Gouse,M.,Kirsten,J.,Jenkins,L.(2003).BtcottoninSouthAfrica:adoptionandtheimpactonfarmincomesamongstsmall-scaleandlarge-scalefarmers. Agrekon42:15-28.

Gouse,M.,Kirsten,J.,Shankar,B.,Thirtle,C.2005.BtcottoninKwazuluNatal:technologicaltriumphbutinstitutionalfailure.AgBiotechNet7:1-7.

Govereh,J.,Jayne,T.(1999).Effectsofcashcropproduction onfoodcropproductivityinZimbabwe:synergiesortrade-offs?MSU International Development Working Paper N°. 74,EastLansing,MI:MichiganStateUniversity.

Heuberger,S.,Ellers-Kirk,C.,Yafuso,C.,Gassmann,A.J.,Tabashnik,B.E.,Dennehy,T.J,Carrière,Y.(2008).EffectsofrefugecontaminationbytransgenesonBtresistanceinpinkbollworm(Lepidoptera:Gelechiidae). Journal of Economic Entomology101:504-514.

Höfte,H.,Whiteley,H.R.(1989).InsecticidalcrystalproteinsofBacillus thuringiensis. Microbiology Review53:242-255.

ICAC(2017).Cotton:worldstatistics.International Cotton Advisory Committee, Washington DC.

ISAAA(2016).GlobalStatusofCommercializedBiotech/GMCrops:2016. ISAAA Brief N°. 52.ISAAA:Ithaca,NY.

Isaacman,A.,Roberts,R.(1995).Cotton,colonialism,andsocialhistoryinSub-SaharanAfrica.London:JamesCurrey.

Ismael,Y.,Bennet,R.,Morse,S.(2002).BenefitsfromBtcottonusebysmallholderfarmersinSouthAfrica. AgBioForum5(1):1-5.

James,C.(2015).GlobalstatusofcommercializedBiotech/GMcrops:2015.ISAAA Brief No 51ISAAA:Ithaca,NY.

Jensen,N.,Barrett,C.(2017).Agriculturalindexinsurancefordevelopment.Applied Economic Perspectives and Policy39(2):199-219.

Kirsten,J.,Gouse,M.(2003).TheadoptionandimpactofagriculturalbiotechnologyinSouthAfrica.InKalaitzandonakesN.(ed.)The Economic and Environmental Impacts of Agbiotech,NewYork:KluwerAcademic/PlenumPublishers,pp.243-259.

Kooistra,K.,Pyburn,R.,Termorshuizen,A.(2006).TheSustainabilityofcotton.Consequencesformanandtheenvironment.ScienceShopWageningenUniversityandResearchCenterReportN°.223.

Kranthi,K.,Naidu,S.,Dhawad,C.,Tatwawadi,A.,Mate,K., Patil,E.,Bharose,A.,Behere,G.Wadaskar,R., Kranthi,S.(2005).Temporalandintra-plantvariabilityofCry1AcexpressioninBt-cottonanditsinfluenceonthesurvivalofthecottonbollworm,Helicoverpa armigera (Hübner)(Noctuidae:Lepidoptera). Current Science89(2):291-298.

Krishna,V.V.,Qaim,M.(2012).Btcottonandsustainability ofpesticidereductionsinIndia. Agricultural Systems 107:47-55.

Krishna,V.,Qaim,M.,Zilberman,D.(2016).Transgeniccrops,productionriskandagrobiodiversity. Eur Rev Agric Econ43(1):137-164.

7

49

Liu,Y.,GeF.,Liang,Y.,Wu,G.,Li,J.(2015).Characterization ofcompetitiveinteractionsinthecoexistenceof Bt-transgenicandconventionalrice. BMC Biotechnology 15(27):1-12.

Lu,Y.,Wu,K.,Jiang,Y.,Guo,Y.,Desneux,N.(2012).WidespreadadoptionofBtcottonandinsecticidedecreasepromotesbiocontrolservices.Nature487:362-367.

Men,X.Y.,Ge,F.,Edwards,C.A.,Yardim,E.N.(2005). TheinfluenceofpesticideapplicationsonHelicoverpa armigera HübnerandsuckingpestsintransgenicBtcottonandnon-transgeniccottoninChina. Crop Protection24:319-24.

Meng,E.C.H.,Hu,R.,Shi,X.,Zhang,S.(2006).MaizeinChina:productionsystems,constraintsandresearchpriorities.Mexico,D.F.:CIMMYT.

Minot,N.,Daniels,L.(2005).Impactofglobalcotton marketsonruralpovertyinBenin. Agricultural Economics33:453-466.

Morse,S.,Bennett,R.,Ismael,Y.(2004).WhyBtcotton paysforsmall-scaleproducersinSouthAfrica. Nature Biotechnology22(4):379-380.

Naylor,R.(1994).HerbicideuseinAsianriceproduction. World Development22:55-70.

OECD(2006).CottoninWestAfrica.The Economic and Social Stakes,Paris:OECD.

Oerke,E.C.(2006).Croplossestopests. J. Agric. Sci.144(1):31-43.

Pemsl,D.(2006).EconomicsofAgriculturalBiotechnologyinCropProtectioninDevelopingCountries–TheCaseofBt-CottoninShandongProvince,China.Pesticide Policy Project Publication SeriesNo.11,Hannover,Germany:UniversityofHannover.

Pertry,I.,Sanou,E.I.R.,Speelman,S.,Ingelbrecht,I.(2016). ThesuccessstoryofBtcottoninBurkinaFaso:arolemodelforsustainablecottonproductioninothercotton-growingcountries?In:Innovative farming and forestry across the emerging world, the role of genetically modified crops and trees,DeBuckS.,IngelbrechtI.,HeijdeM. andVanMontaguM.(eds.).

Potrykus,I.,Bilang,R.,Futterer,J.,Sautter,C.,Schrott,M.(1998).Geneticengineeringofcropplants.InAltmanA.(ed.)Agricultural Biotechnology,NewYork:MarcelDekker, Inc,pp.119-159.

Pray,C.,Huang,J.,Hu,R.,Rozelle,S.(2002).FiveyearsofBtcottoninChina–thebenefitscontinue. The Plant Journal 31:423-430.

Putterman,L.(1995).EconomicreformandsmallholderagricultureinTanzania:adiscussionofrecentmarketliberalization,roadrehabilitation,andtechnologydisseminationefforts.WorldDevelopment23(2):311-326.

Qaim,M.(2003).BtcottoninIndia:Fieldtrialresultsandeconomicprojections.World Development31:2115-2127.

Qaim,M.,deJanvry,A.(2003).Geneticallymodifiedcrops,corporatepricingstrategies,andfarmers’adoption: thecaseofBtcottoninArgentina.Amer. J. Agr. Econ.85(4):814-828.

Qaim,M.,deJanvry,A.(2005).BtcottonandpesticideuseinArgentina:economicandenvironmentaleffects.Environment and Development Economics10:179-200.

Qaim,M.,Subramanian,S.,Naik,G.,Zilberman,D.(2006).AdoptionofBtcottonandimpactvariability:insightsfromIndia.Review of Agricultural Economics 28:48-58.

Qaim,M.,Zilberman,D.(2003).Yieldeffectsofgeneticallymodifiedcropsindevelopingcountries. Science299:900-902.

Qiao,F.(2015).FifteenyearsofBtcottoninChina: theeconomicimpactanditsdynamics. World Development70:177-185.

Sahn,D.(1990).Theimpactofexportcropproduction onnutritionalstatusinCôted’Ivoire. World Development18:1635-1653.

Schnurr,M.A.(2012).InventingMakhathini:creatingaprototypeforthedisseminationofGeneticallyModifiedcropsintoAfrica.Geoforum43(4):784–792.

Shankar,B.,Thirtle,C.(2005).Pesticideproductivityandtransgeniccottontechnology:TheSouthAfricansmallholdercase.Journal of Agricultural Economics56:97-116.

Sidhu,M.S.(1999).ImpactofintellectualpropertyrightsontheIndianseedindustry.Indian Journal of Agricultural Economics54:370-379.

Skinner,D.Z.,Muthukrishnan,S.,Liang,G.H.(2004).Transformation:apowerfultoolforcropimprovement.InLiangG.H.andSkinnerD.Z.(eds),Genetically Modified Crops: Their Development, Uses, and Risks,NewYork: FoodProductsPress,pp1-16.

Smith,R.H.,Smith,J.W.,Park,S.H.(2004).Cottontransformation:successesandchallenges.InLiangG.H.andSkinnerD.Z.(eds)Genetically Modified Crops: their Development, uses and risks, NewYork:FoodProductsPress,pp.247-257.

Stoll,S.(2002).LardingtheLeanEarth.SoilandSocietyinNineteenthCenturyAmerica.NewYork:HillandWang.

Stone,G.(2004).BiotechnologyandthepoliticalecologyofinformationinIndia.Human Organisation63:127-140.

Stone,G.(2007).AgriculturaldeskillingandthespreadofgeneticallymodifiedcottoninWarangal. Current Anthropology48:67-103.

Tabashnik,B.E.(1994).Delayinginsectadaptationtotransgenicplants:seedmixturesandrefugiareconsidered.Proceedings of the Royal Society London255:7-12.

Tabashnik,B.E.,Brévault,T.,Carrière,Y.,2013.InsectresistancetoBtcrops:lessonsfromthefirstbillionacres.Nature Biotechnology31(6):510-521.

TechnoServeandIFAD(2011).Technicalbriefonoutgrowerschemes–Enhancingprofitability.http://www.technoserve.org/resources.

Theriault,V.,Serra,R.(2014).Institutionalenvironment andtechnicalefficiency:astochasticfrontieranalysis ofcottonproducersinWestAfrica. Journal of Agricultural Economics65:383-405.

Theriault,V.,Tschirley,D.L.(2014).Howinstitutionsmediatetheimpactofcashcroppingonfoodcropintensification:anapplicationtocottoninSub-SaharanAfrica. World Development64:298-310.

Thompson,W.M.,Anderson,J.,Ibendahl,G.,Hudson,D.(2007).TheimpactsofGMseedtechnologyoncotton:costofproductioninMississippi,1996-2005.PaperpresentedattheSouthernAgriculturalEconomicsAssociationAnnualMeeting,Mobile,Alabama,February2007.

Traoré,O,Denys,S,Vitale,J,Traoré,K,Bazoumana,K(2008).Economicsandmarketing:testingtheefficacyandeconomicpotentialofBollgardIIunderBurkinaFasocroppingconditions. J Cotton Sci 12:87–98.

Tripp,R.(2009).Biotechnologyandagriculturaldevelopment.Transgeniccotton,ruralinstitutionsandresource-poorfarmers.ByTrippR.(ed.)LondonandNewYork:Routledge.

Tschirley,D.,Poulton,C.,Gergely,N.,Labaste,P.,Baffes,J.,Boughton,D.,Estur,G.(2010).Institutionaldiversity andperformanceinAfricancottonsectors. Development Policy Review 28(3):295-323.

Cotton in Africa

Tschirley,D.,Poulton,C.,Labaste,P.(2009).OrganizationandperformanceofcottonsectorinAfrica:learningfromreformexperience.World Bank,Washington,DC.

VIBFactsSeries(2013).BtcottoninIndia–Asuccessstoryfortheenvironmentandlocalwelfare.

VIBFactsSeries(2016).Fromplanttocrop:thepast,presentandfutureofplantbreeding.

Vitale,J.,Greenplate,J.(2014).Theroleofbiotechnologyinsustainableagricultureofthetwenty-firstcentury:thecommercialintroductionofBollgardIIinBurkinaFaso.In Convergence of Food Security, Energy Security and Sustainable Agriculture, Biotechnology in Agriculture and Forestry,Vol.67,D.D.Songstad,J.L.Hatfield,and D.T.Tomes(Eds.).Berlin,Springer-Verlag,pp.239-293.

Vitale,J.,Ouattarra,M.,Vognan,G.(2011).EnhancingsustainabilityofcottonproductionsystemsinWestAfrica:asummaryofempiricalevidencefromBurkinaFaso.Sustainability3:1136-1169.

Vitale,J.,Vognan,G.,Vitale,P.P.(2016).TheSocio-economicimpactsofGMcottoninBurkinaFaso:doesfarmstructureaffecthowbenefitsaredistributed?AgBioForum19(2):120-135.

Wendimu,M.A.(2016).Jatrophapotentialonmarginal landinEthiopia:realityormyth?Energy for Sustainable Development 30:14–20.

Wu,K.,Li,W.,Feng,H.,Guo,Y.(2002).Seasonalabundanceofthemirids,Lygus lucorumandAdelphocorisspp.(Hemiptera:Miridae)onBtcottoninnorthernChina.Crop Protection21:997-1002.

Zhang,B.H.,Liu,F.,Yao,C.B.,Wang,K.B.(2000).RecentprogressincottonbiotechnologyandgeneticengineeringinChina.Current Science79:37-44.

Zilberman,D.,Ameden,H.,Qaim,M.(2007).Theimpactofagriculturalbiotechnologyonyields,risks,andbiodiversityinlow-incomecountries.Journal of Development Studies 43(1):63-78.

51

IPBO(InternationalPlantBiotechnologyOutreach),whichformspartoftheVIB,wassetupin2000byProf.Em.MarcVanMontaguandGhentUniversity. IPBO’smission is topromote knowledgeand technology transfer in theareaofplantbiotechnologytodevelopingcountries,withafocusonagreenandsustainableagricultureandagro-industry.Toaccomplishthismission,IPBOfocusesoncommunication,traininginplantbreeding,greenbiotechnologyandrelatedbiosafety,andfostersnetworkingandprojectdevelopmenttoimplementcooperationbetweendevelopingcountriesandFlanders.Formoreinformation,visit:www.ipbo.vib-ugent.be

Basic research in life sciences isVIB’s raisond’être.VIB is an independent research institutewhere some1,500 topscientistsfromBelgiumandabroadconductpioneeringbasicresearch.Assuch,theyarepushingtheboundariesofwhatweknowaboutmolecularmechanismsandhowtheyrulelivingorganismssuchashumanbeings,animals,plantsandmicroorganisms.BasedonaclosepartnershipwithfiveFlemishuniversities–GhentUniversity,KULeuven,UniversityofAntwerp,VrijeUniversiteitBrusselandHasseltUniversity–andsupportedbyasolidfundingprogram,VIBunitestheexpertiseofallitscollaboratorsandresearchgroupsinasingleinstitute.TheVIB-UGentCenterforPlantSystemsBiologywantstogaininsightintohowplantsgrowandrespondtotheenvironment.Scientistsstudyhowleavesandrootsareformed,whichmicro-organismsliveonandaroundtheplantandwhichsubstancestheplantmakes.Thisknowledgecanleadtosustainableinnovationsinagricultureandfood.

Aftermorethantwentyyearsofuninterruptedgrowth,Ghent UniversityisnowoneofthemostimportantinstitutionsofhighereducationandresearchintheLowCountries.GhentUniversityyearlyattractsover41,000students,withaforeignstudentpopulationofover2,200EUandnon-EUcitizens.GhentUniversityoffersabroadrangeofstudyprogramsinallacademicandscientificfields.Withaviewtocooperationinresearchandcommunityservice,numerousresearchgroups,centersandinstituteshavebeenfoundedovertheyears.Formoreinformation,visit:www.UGent.be.

The International IndustrialBiotechnologyNetwork (IIBN)wasestablished in2010as a joint initiativeofUNIDO, theFlemishGovernment(EWI)andIPBO.IIBNservesasacatalystforadvancingsustainableapplicationsofagriculturalandindustrialbiotechnology indevelopingandemergingeconomies incooperationwithFlandersandother internationalpartners. IIBN isbeingdevelopedalongthreetracks: (1)engage inadvocacytoraiseawareness for thedevelopmentpotentialofagriculturalbiotechnologybyprovidingscience-basedinformationandcasestudies;(2)establishaformalnetworkoflike-mindedinstitutionsandorganizations,and(3)fosterR&Dcooperationandcapacitybuildinginbiosciencesthat addresses the needs of developing and emerging economies, in cooperation with stakeholders in Flanders andbeyond.

ThedepartmentofEconomy,ScienceandInnovation(EWI Department)oftheFlemishGovernmentprepares,monitorsandevaluatespolicyintheeconomy,scienceandinnovationpolicyarea.TheaimistodevelopFlandersintooneofthemostadvancedandprosperousregionsoftheworld.Theirdrivingforcesarethepromotionof(1)excellenceinscientificresearch,(2)anattractiveandsustainablebusinessstrategyand(3)acreative,innovativeandentrepreneurialsociety.

The United Nations Industrial Development Organization (UNIDO) aims to eradicate poverty through inclusive andsustainableindustrialdevelopment(ISID).UNIDOadvocatesthatISIDisthekeydriverforthesuccessfulintegrationoftheeconomic,socialandenvironmentaldimensions,requiredtofullyrealizesustainabledevelopmentforthebenefitoffuturegenerations.

R.E.SilviaTravella(IPBO/VIB)D/2017/12.267/7

IPBOTechnologiepark3,9052Ghent,BelgiumTel:+3292648727-ipbo@vib-ugent.be-www.ipbo.vib-ugent.be